docs(P02): complete messaging domain phase — v0.4

---ci--- project: atelier phase: 2 milestone: v0.4 status: complete phase_role: execution phase_tag: v0.3.2 requirements: covered: [ATELIER-97, ATELIER-98, ATELIER-99, ATELIER-100, ATELIER-101] partial: [] ---/ci---
docs(ship): P1 complete — v0.3.1 tagged, release 479
2026-08-05 16:00:31 +00:00 · 2026-08-05 15:53:16 +00:00 · 2026-08-05 15:51:03 +00:00 · 2026-08-05 14:03:35 +00:00 · 2026-08-05 14:01:13 +00:00 · 2026-08-05 03:45:38 +00:00
41 changed files with 7743 additions and 80 deletions
@@ -1,11 +1,15 @@
 {
  "phase": 1,
  "stage": "complete",
-  "milestone": "v0.3",
+  "milestone": "v0.4",
  "phase_role": "execution",
  "project": "atelier",
  "attempts": 0,
-  "updated_at": "2026-08-05T03:25:00Z",
+  "updated_at": "2026-08-05T06:15:00Z",
-  "phase_tag": "v0.2.1",
+  "milestone_complete": false,
-  "release_id": 469
+  "milestone_branch": "milestone/v0.4-edge-quantum-langs",
  "phase_branch": "phase/01-edge",
  "tag_base": "v0.3",
  "phase_tag": "v0.3.1",
  "release_id": 479
 }
@@ -0,0 +1,100 @@
 # Atelier v0.4 — Clarify Stage
 > Generated during v0.4 P0 CLARIFY stage. Autonomy level: `full` (defaults auto-accepted).
 > All decisions below are above the 0.60 confidence threshold → auto-decided, no HITL.
 > Decisions continue from D-060 (last v0.3 decision). Scope: NEW v0.4 work only (edge, messaging, language-derived docs). Prior milestone decisions (D-001..D-060) are not re-clarified.
 ## Ambiguities Resolved: 9
 ### 1. Edge domain P-rules — boundary vs performance/concurrency/observability (avoid duplication)
 - **Question:** The edge domain is latency-bound, resource-constrained, and geo-distributed — concerns that overlap with the existing performance domain (caching P5, lazy P6, async P7), concurrency domain (bounded queues P9, immutability P1), and observability domain (structured logs, correlation). How do edge's 10 P-rules distinguish themselves without duplicating existing domain P-rules?
 - **Default resolution:** Edge owns the **proximity/location angle**: compute placed near the user or data source, location-aware routing, resource constraints of edge nodes, and offline-capability. The boundary is: performance owns *generic measurement & optimization*; concurrency owns *in-process synchronization primitives*; observability owns *generic logging/metrics/tracing*; edge owns *the placement, constraint, and disconnection concerns that only arise at the network edge*. Edge P-rules trace heavily to C4 Locality (the defining edge trait — compute near data/user) and C8 Economy (constrained devices), plus C5 Reversibility (offline/sync reversibility), C7 Observability (geo-distributed visibility), C1 Correctness (eventual-consistency correctness under partition). Edge is a broad-derivation domain (6–8 C-rules, parallel to i18n/gitops). Cross-links to performance/concurrency/observability are one-directional outward (per D-026 extended); no back-link edits.
 - **Rationale:** Every prior domain owned a distinct "angle" on shared core principles (k8s P8 storage-explicit vs iac P3 state-is-truth — same C-rules, different scope per D-026/RESEARCH). Edge's unique contribution is *where compute runs*, which no existing domain addresses. The C4 Locality emphasis is the discriminator: performance's locality is algorithmic (data near compute); edge's locality is geographic (compute near user). Confirmed by RESEARCH.md pattern: "each domain owns its own principle angle."
 - **Confidence:** 0.88
 - **Alternatives rejected:** (a) Edge as a derived doc under performance — rejected: geo-distribution + offline + IoT constrained-device concerns warrant a full 10-P-rule set, not a single derived doc (parallel to D-020 k8s-vs-devops reasoning); (b) Edge reusing performance P5/P6/P7 verbatim — rejected: would orphan edge rules from their actual derivation (C4 Locality, C8 Economy) and violate the "no duplicate rows" matrix invariant.
 - **Decision:** D-061
 ### 2. Messaging domain boundary vs concurrency/patterns.md (in-process vs cross-system)
 - **Question:** `domains/concurrency/patterns.md` already covers "Message Passing" (Pattern 1), "Bounded Queue with Backpressure" (Pattern 5), and "Actor Model" (Pattern 3). The new messaging domain covers queues, pub/sub, streams, and delivery semantics. Where is the boundary — does messaging duplicate concurrency's queue/message-passing patterns?
 - **Default resolution:** Concurrency owns **in-process** synchronization primitives (threads, locks, channels, async within one program). Messaging owns **cross-process / cross-system** asynchronous communication (brokers: Kafka, RabbitMQ, SQS, NATS, Pulsar). The discriminator: messaging concerns arise *only when producer and consumer are separate systems* — delivery semantics (at-least-once/at-most-once/exactly-once), ordering across partitions, idempotency across retries, dead-letter queues, replay from offsets, consumer-group coordination. Concurrency's "bounded queue" is the in-process analog; messaging's "bounded queue" is the broker-backed analog with network-failure semantics. Messaging P-rules trace to C1 Correctness (delivery guarantees), C5 Reversibility (replay/rollback), C7 Observability (lag, DLQ visibility), C6 Composability (pub/sub decoupling). Cross-link concurrency/patterns.md ← messaging (one-directional); messaging references concurrency as the in-process foundation but owns the distributed-delivery layer above it.
 - **Rationale:** The v0.2 RESEARCH established the precedent: gitops-operators is "the deployment-automation layer above kubernetes/infrastructure-as-code — borrows their model, adds the git-as-source-of-truth layer." Messaging is the same: the async-delivery layer above concurrency — borrows the bounded-queue/backpressure model, adds cross-system delivery/ordering/idempotency. Concurrency/patterns.md Pattern 5 explicitly scopes itself to in-process ("producer is blocked or signaled"); messaging extends this to network-partition scenarios. No duplication because the *failure model* differs (crash vs network failure).
 - **Confidence:** 0.86
 - **Alternatives rejected:** (a) Messaging as a derived doc under concurrency — rejected: delivery semantics, stream processing, and broker comparison warrant a full 10-P-rule domain (parallel to D-021 gitops-unification reasoning: splitting would fragment P-rules); (b) Merge messaging P-rules into concurrency — rejected: would bloat concurrency to 20 rules and conflate in-process vs distributed concerns, breaking the "10 P-rules per domain" invariant (D-018/D-030); (c) No cross-link to concurrency — rejected: the bounded-queue/backpressure lineage should be explicit (one outbound cross-link per derived doc per ATELIER-114).
 - **Decision:** D-062
 ### 3. Language-derived docs structure — no new P-rules, section-header trace pattern (confirm D-059)
 - **Question:** PROJECT.md states language-derived docs "trace to existing domain P-rules, not new matrix rows" (the v0.4 Cross-Domain Integration bullet). But how exactly is a language-derived doc structured? The existing `languages/*.md` stubs use section headers like `## Type System (C1 Correctness, Data P7 Type Fidelity)` — they have NO numbered P1–P10 rules. Do the 16 new derived docs (ts-types, ts-tooling, etc.) introduce their own P-rules, or follow the same section-header-trace pattern with no new P-rule numbering?
 - **Default resolution:** Language-derived docs introduce **NO new P-rules**. Each derived doc follows the existing stub structure: section headers that name the traced core C-rule(s) and domain P-rule(s) inline (e.g., `## Strict Null Checks (C1 Correctness, Errors P1 Errors are Data)`), with bullet-point guidance and fenced code. The 4 derived docs per language each focus on one facet (types, tooling, async/concurrency, testing) and trace to the relevant existing domain P-rules. The first-principles stub for each language (typescript.md, etc.) is *extended* with a pointer to its derived docs but retains its existing section structure — it does NOT gain P1–P10 numbering. Matrix row count stays at 190 (170 + 20 for edge+messaging only).
 - **Rationale:** ARCHITECTURE.md dependency direction states `languages/` depend on `domains/` ("language rules apply domain rules") — languages are an *application lens*, not a principle source. The existing stubs confirm this: they trace to `Data P7`, `Errors P1`, `Concurrency P7`, `Observability P1`, `DevOps P2` — never inventing new P-rules. D-059 (referenced in PROJECT.md v0.4 section) already fixed the matrix treatment; this decision confirms the *document structure* that implements it. Adding P1–P10 to languages would break the hierarchy (languages would become peer domains, violating "conflicts go UP, never sideways" and the core>domains>languages ordering).
 - **Confidence:** 0.92
 - **Alternatives rejected:** (a) Language-derived docs get their own P1–P10 per language (40 new P-rules) — rejected: violates ARCHITECTURE.md hierarchy (languages apply, not derive) and D-059's matrix ruling; would also push matrix to 230 rows, breaking the stated 190 total; (b) Hybrid: first-principles stub gains P-rules, derived docs don't — rejected: inconsistent within a language and still introduces new matrix rows contrary to PROJECT.md; (c) Derived docs are pure code examples with no principle traces — rejected: would orphan them from the matrix/traceability contract and lose the "why over what" tech-writer constraint.
 - **Decision:** D-063
 ### 4. Language-derived doc file naming — ts-/py-/go-/rs- prefixes confirmed
 - **Question:** PROJECT.md v0.4 uses short prefixes for derived doc filenames: `ts-types.md`, `py-types.md`, `go-types.md`, `rs-ownership.md` — but the parent first-principles files use full names (`typescript.md`, `python.md`, `go.md`, `rust.md`). The prefix for Rust is `rs-` (not `rust-`), and Go uses `go-` (matching its parent). Is the ts-/py-/go-/rs- prefix scheme correct, or should it be revisited for consistency (e.g., typescript-types, or rust-ownership)?
 - **Default resolution:** Confirm the **ts-/py-/go-/rs-** prefix scheme as specified in PROJECT.md. `ts-` (TypeScript), `py-` (Python), `go-` (Go), `rs-` (Rust). These match the dominant ecosystem conventions: `ts` is the universal TypeScript abbreviation (tsconfig, ts-node, .ts extension); `py` is the universal Python abbreviation (.py, pyproject, py.test); `go` matches the language's own binary/module naming; `rs-` is the established Rust ecosystem prefix (crate names, file conventions). The full parent filename + short-prefix-derived-doc pattern mirrors how domains work (`kubernetes/` parent + `workloads.md` derived — the derived doc doesn't repeat the full domain name).
 - **Rationale:** Consistency with domain-derived-doc naming: derived docs use short, distinct slugs (kubernetes/workloads.md, not kubernetes-kubernetes-workloads.md). The short prefix disambiguates the language while keeping filenames compact. `rs-` over `rust-` follows the broader Rust ecosystem (the `rs-` prefix is more common in registry/package naming than `rust-`, which can collide with the `rust` toolchain). `go-` is correct because Go's ecosystem uses `go` universally and the parent file is already `go.md`.
 - **Confidence:** 0.84
 - **Alternatives rejected:** (a) Full-name prefixes (typescript-types.md, python-types.md, rust-ownership.md) — rejected: verbose, inconsistent with domain-derived-doc compactness, and `rust-` collides with toolchain naming conventions; (b) Single-letter or extension-based (ts/types.md subdirectory per language) — rejected: changes the flat `languages/` structure established in v0.1 (MANIFEST lists `languages/typescript.md`, not `languages/typescript/first-principles.md`); introduces directory nesting the v0.1 manifest schema doesn't support; (c) Normalize Rust to `rust-` for readability — rejected: breaks ecosystem convention alignment; `rs-` is the recognized Rust short prefix.
 - **Decision:** D-064
 ### 5. Language-derived docs in MANIFEST — new languages/ section (closes pre-existing drift)
 - **Question:** The current `MANIFEST.md` (post-v0.3) has sections for Core, Domains, Examples, and Cross-Cutting — but **no `languages/` section at all**, despite `languages/{typescript,python,go,rust}.md` existing since v0.1 (ATELIER-31, P6). This is pre-existing manifest drift (parallel to the v0.2 ESC-002 note where `examples/` was unlisted). Where do the 16 new language-derived docs (and the 4 existing stubs) go in the manifest?
 - **Default resolution:** Add a **new top-level `## Languages` section** to `MANIFEST.md` in P4, parallel to the `## Domains` and `## Examples` sections. The section lists each language (TypeScript, Python, Go, Rust) with its first-principles stub (✓) and its 4 derived docs. This *both* lists the 16 new v0.4 derived docs AND closes the pre-existing drift by listing the 4 v0.1 stubs that were never added. The section appears after `## Domains` and before `## Examples` (matching the reading order: core → domains → languages). The manifest's Cross-Cutting summary line updates: "19 domains, 190 P-rules post-v0.4" and notes that languages apply (not derive) domain rules.
 - **Rationale:** The manifest is authoritative — "unlisted = not part of the framework" (PROJECT.md constraint, MANIFEST.md header). The v0.2 audit escalation ESC-002 established the precedent: when a directory exists but is unlisted, the next milestone that touches it adds the listing (IDEATE-17 → ATELIER-91 closed examples/ drift in v0.3). v0.4 touches `languages/` for the first time since v0.1, so it closes the languages/ drift on the same principle. Placing the section between Domains and Examples matches `core/reading-order.md` (languages are read after domains, before examples) and the ARCHITECTURE.md dependency direction (languages depend on domains). This is a single P4 task, not split across phases.
 - **Confidence:** 0.90
 - **Alternatives rejected:** (a) List language-derived docs under the existing `## Domains` section — rejected: languages are not domains (ARCHITECTURE.md distinguishes the two levels; mixing them breaks the hierarchy and the matrix row-count invariant); (b) Defer the 4 v0.1 stubs to a future milestone, list only the 16 new docs — rejected: would leave the pre-existing drift half-closed and is inconsistent with the ESC-002 precedent (close the whole directory when you touch it); (c) Put languages/ at the end after Cross-Cutting — rejected: violates reading order (languages are consumed before examples/matrix reference, per core/reading-order.md).
 - **Decision:** D-065
 ### 6. Matrix treatment of language-derived docs — NO new matrix rows; traceability via domain-coverage sub-table + inline traces
 - **Question:** D-059 (PROJECT.md) says language-derived docs "trace to existing domain P-rules, not new matrix rows." But traceability must be *captured somewhere* — where? Does `matrix/principles-matrix.md` gain a languages section (with non-numbered trace rows)? Does `matrix/domain-coverage.md` gain a languages sub-table? How is "this language-derived doc traces to these domain P-rules" recorded verifiably?
 - **Default resolution:** `matrix/principles-matrix.md` gains **NO language rows** — it stays at 190 P-rules (170 + 20 for edge+messaging), 19 domains. `matrix/domain-coverage.md` gains a **new `## Languages Coverage` sub-table** (per ATELIER-107) with schema: `Language | Derived Doc | Traced Domain P-rules | Traced Core C-rules`. Each of the 16 derived docs gets one row listing the domain P-rules its section-headers trace to (e.g., `TypeScript | ts-types.md | Data P7 Type Fidelity, Errors P1, C1 | C1, C6`). Traceability is ALSO captured inline in each derived doc's section headers (the existing stub pattern: `## Section (C-rule, Domain P-rule)`). The domain-coverage sub-table is the verifiable aggregate; the inline headers are the per-doc detail. The agent-checklist (ATELIER-108) gains a language-derived-docs check: "every derived-doc section header names ≥1 traced domain P-rule."
 - **Rationale:** The principles-matrix is the domain→core join table; languages are not domains and have no P-rules to map (D-063), so adding rows would either be empty or fake. The domain-coverage sub-table is the right home: it already holds the per-domain coverage schema (IDEATE-03) and the C-rule→domains table (IDEATE-19) — a languages sub-table is the natural extension, parallel in shape. This gives reviewers a single place to verify "no language-derived doc is orphaned from its traced domain P-rules" without scanning all 16 files. Mirrors how v0.3's C-rule coverage table (IDEATE-19) was the complement to the per-domain row schema.
 - **Confidence:** 0.87
 - **Alternatives rejected:** (a) Add a `## Languages` section to principles-matrix.md with one row per derived doc → C-rule — rejected: the matrix maps P-rules, and languages have no P-rules (D-063); this would conflate the join table's semantics and inflate the "190 P-rules" count that PROJECT.md/ATELIER-106 explicitly state; (b) Traceability only inline in section headers, no aggregate table — rejected: not verifiable in one place; reviewers would have to open 16 files to confirm no orphans, violating the domain-expert persona's "no orphans in matrix" constraint; (c) Add languages as a 20th "domain" row in the per-domain coverage table — rejected: languages are not domains (ARCHITECTURE.md hierarchy); would corrupt the domain count invariant.
 - **Decision:** D-066
 ### 7. Phase-specific personas — edge-engineer (P1) + languages-engineer (P3); messaging (P2) by tech-writer + domain-expert with edge-engineer consultation
 - **Question:** ROADMAP v0.4 names "edge-engineer + languages-engineer phase-specific personas." That is only TWO named personas for three execution phases (P1 edge, P2 messaging, P3 languages). Does the messaging domain (P2) get its own phase-specific persona (e.g., messaging-engineer), or is it covered by existing/adjacent personas? If the latter, who owns messaging P2 authorship and with what constraints?
 - **Default resolution:** Add **two** phase-specific personas: `edge-engineer` (active P1, removed after milestone) and `languages-engineer` (active P3, removed after milestone). The **messaging domain (P2) is authored by tech-writer + domain-expert** (the permanent roster), with edge-engineer consulted for the distributed-systems/async overlap (D-062 boundary). edge-engineer constraints: ["latency is a correctness constraint", "compute placement is a design decision", "offline is a first-class state", "trace to core", "10 P-rules per domain", "no runtime code", "resource constraints are explicit", "sync conflict is a correctness failure"]. languages-engineer constraints: ["apply, don't derive (D-063)", "trace every section to an existing domain P-rule", "language idioms serve principles, not the reverse", "no new P-rules", "no new matrix rows (D-066)", "fenced code is illustrative only"]. Messaging P2 gets a task must-have baked into the plan: "P1/P2 demonstrated edge + messaging boundary (D-062); delivery-semantics comparison table required (D-069)" so the edge-engineer's distributed-systems lens survives the P1→P2 persona transition (parallel to D-051's ml-engineer-constraints-baked-into-task pattern from v0.3).
 - **Rationale:** The ROADMAP explicitly names only 2 personas — this is the authoritative signal (the ROADMAP is finalized in PLAN and pre-named personas per PROJECT.md v0.4 section + ROADMAP P0 row). v0.3 set the precedent (D-022): smaller-surface or within-competence domains don't get new personas (i18n + compliance in v0.3 P3 used tech-writer + domain-expert, no new personas). Messaging, while specialized, is within domain-expert's cross-cutting competence (delivery semantics, idempotency, ordering are correctness/observability derivations the domain-expert already validates). Limiting to 2 phase-specific personas holds the active roster to 5 (3 permanent + 2 phase-specific), matching v0.3's peak and avoiding the "persona explosion" risk flagged in v0.3 RESEARCH. The edge-engineer consultation for P2 ensures the D-062 in-process-vs-cross-system boundary is respected without a 3rd persona.
 - **Confidence:** 0.80
 - **Alternatives rejected:** (a) Three phase-specific personas (edge-engineer P1, messaging-engineer P2, languages-engineer P3) — rejected: ROADMAP names only 2; would push roster to 6 active (3+3), exceeding v0.3's 5-persona peak and raising the explosion risk; messaging's specialty, while real, is less novel than ai-ml's (which warranted ml-engineer per D-019/D-020); (b) One persona (edge-engineer) covers P1+P2, languages-engineer P3 — rejected: edge (latency/geo/resource) and messaging (delivery/ordering/idempotency) are distinct specialties; forcing one persona blurs the D-062 boundary and dilutes edge-specific constraints; (c) No new personas, all three phases by tech-writer + domain-expert — rejected: edge's latency-as-correctness and compute-placement concerns, and languages' apply-don't-derive discipline, benefit from specialist constraints the generalist personas lack (v0.2 D-014 / v0.3 D-019 precedent).
 - **Decision:** D-067
 ### 8. Examples — confirm 2 good + 2 bad pre-named set; bad examples cite breached P-rules
 - **Question:** PROJECT.md / ATELIER-112/113 pre-name the 4 v0.4 examples: `good/edge-offline-sync.md`, `good/messaging-idempotent-consumer.md`, `bad/edge-sync-loop.md`, `bad/messaging-shared-subscription.md`. Are these the right picks (vs alternatives like good/cdn-cache-invalidation or bad/messaging-blocking-consumer), and what breach semantics must the bad examples follow?
 - **Default resolution:** Confirm the pre-named set as specified. **Good:** `edge-offline-sync.md` (demonstrates offline-first + sync conflict resolution via CRDT or LWW — the highest-illustration-value edge case, exercising P1 offline + the sync.md decision matrix), `messaging-idempotent-consumer.md` (demonstrates exactly-once-via-idempotency-keys + dedup — the highest-illustration-value messaging case, exercising delivery-semantics + idempotency). **Bad:** `edge-sync-loop.md` (a sync conflict that oscillates/loops forever — the chaos anti-pattern pre-specified in ATELIER-110), `messaging-shared-subscription.md` (a shared subscription causing duplicate processing / fan-out duplication — the chaos anti-pattern pre-specified in ATELIER-110). Each bad example must cite the **specific P-rule breached** (per D-025 / IDEATE-07 named-bad-example convention carried forward v0.1→v0.2→v0.3). `edge-sync-loop.md` breaches the edge sync P-rule (exact P# TBD in P1, but it is the "sync conflicts are bounded, not infinite" rule); `messaging-shared-subscription.md` breaches the messaging subscription P-rule (exact P# TBD in P2, but it is the "consumer identity is per-consumer, not shared" rule). Both bad examples are single-breach (unlike v0.3's two-breach compliance example D-044 — edge/messaging breaches are cleaner as single-rule violations).
 - **Rationale:** The pre-named set covers one good + one bad per new domain — balanced, parallel to v0.2 (terraform + k8s) and v0.3 (gitops + ai-ml good; i18n + compliance bad). D-025 (v0.2) established 2-good + 2-bad for 2-domain milestones; v0.4 follows the same ratio. The chaos anti-patterns (sync loop, shared subscription) are pre-specified in ATELIER-110 because they are the highest-illustration-value *fundamental* violations — a sync loop is to edge what push-deploy was to GitOps (IDEATE-24); a shared subscription is to messaging what an orphan model was to ai-ml (IDEATE-27). Alternatives (cdn-cache-invalidation good, blocking-consumer bad) are lower-value: cache invalidation is a performance-domain concern (cross-link only), and blocking-consumer overlaps concurrency/patterns.md (D-062 boundary). Single-breach is correct because edge/messaging P-rules are individually clear; the two-breach pattern (D-044) was specific to compliance where two breaches (append-only + redaction) compound the highest-stakes domain.
 - **Confidence:** 0.85
 - **Alternatives rejected:** (a) `good/cdn-cache-invalidation.md` instead of edge-offline-sync — rejected: cache invalidation is primarily a performance/P5 concern; offline-sync exercises more edge-unique P-rules (offline state, sync, conflict); (b) `bad/messaging-blocking-consumer.md` instead of shared-subscription — rejected: blocking-consumer is a concurrency/patterns.md concern (D-062 in-process boundary); shared-subscription is a cross-system messaging-unique failure; (c) Two-breach bad examples (per D-044 compliance pattern) — rejected: D-044 was domain-specific to compliance's compounded stakes; edge/messaging single-rule breaches are clearer and match v0.1/v0.2 bad-example convention; (d) 4 good + 4 bad (one per derived doc) — rejected: would unbalance P5 (D-025 reasoning: 2+2 covers highest-illustration-value; remaining coverage via cross-links + anti-patterns).
 - **Decision:** D-068
 ### 9. Derived-doc decision/comparison tables — confirm queues/sync matrices + add streams platform comparison
 - **Question:** v0.2 established the decision-matrix pattern (Helm vs Kustomize in ATELIER-46/47), v0.3 transferred it to 3 derived docs (ArgoCD vs Flux, DVC vs Delta Lake vs LakeFS, OPA vs Cedar vs Kyverno vs Sentinel). v0.4 REQUIREMENTS pre-specify two: `queues.md` needs an at-least-once/at-most-once/exactly-once comparison (ATELIER-98), and `sync.md` needs a CRDT/last-write-win decision matrix (ATELIER-96). But `streams.md` (ATELIER-100) and `delivery-semantics.md` (ATELIER-101) have no comparison table specified. Should they also get one, following the established pattern?
 - **Default resolution:** Confirm the two pre-specified matrices (queues delivery-semantics comparison in queues.md; CRDT-vs-LWW decision matrix in sync.md). Add **two more** to maintain pattern parity (every v0.4 derived doc that compares distinct tools/models gets a decision/comparison table, per the v0.2→v0.3 transfer convention): `streams.md` gets a **stream-platform comparison table** (Kafka vs Kinesis vs Pulsar/NATS JetStream) covering ordering guarantees, partitioning model, replay/retention, and use-case fit; `delivery-semantics.md` gets a **dead-letter strategy comparison table** (retry-count-limit vs TTL-with-backoff vs poison-queue vs DLQ+alert) covering when each applies, failure visibility, and operational cost. Additionally, `cdn.md` (ATELIER-93) already has a pre-specified "edge-cache vs origin decision matrix" — that is confirmed (the cache-hit/miss/invalidation tradeoff). Total: 4 decision/comparison tables across the 9 new derived docs (cdn, queues, sync, streams, delivery-semantics = 5 candidate docs; 4 get tables — delivery-semantics gets a strategy comparison rather than a tool-vs-tool matrix since it's a pattern comparison, not a product comparison).
 - **Rationale:** The v0.2→v0.3 transfer pattern (REQUIREMENTS v0.3 "Within-Project Pattern Transfer" table, IDEATE-21/22/23) established that derived docs comparing distinct tools/models MUST include a decision/comparison table — this is now a project convention, not a one-off. streams.md compares Kafka/Kinesis/Pulsar (3+ distinct platforms with different ordering/replay models) — directly parallel to v0.3's data-versioning tool comparison (DVC/Delta Lake/LakeFS, IDEATE-22). delivery-semantics.md compares dead-letter strategies (distinct failure-handling models) — parallel to v0.2's Helm-vs-Kustomize "when X fits" guidance. Omitting tables from streams/delivery-semantics would be an inconsistency reviewers should flag (the v0.3 transfer-table verification was a P4 review check). cdn.md's edge-cache-vs-origin matrix is already required (ATELIER-93) and is the cache-placement analog of Helm-vs-Kustomize.
 - **Confidence:** 0.82
 - **Alternatives rejected:** (a) Only the 2 pre-specified tables (queues, sync), none for streams/delivery-semantics — rejected: breaks the v0.2→v0.3 transfer convention; streams.md comparing 3 platforms without a table would be the first multi-tool derived doc since v0.2 to omit one; (b) Tables in all 5 candidate docs including a tool-vs-tool table in delivery-semantics — rejected: delivery-semantics compares *strategies/patterns*, not products; forcing a product table (e.g., "Kafka DLQ vs RabbitMQ DLQ") would be misleading since DLQ semantics are cross-platform; a strategy comparison is the correct shape; (c) Defer streams/delivery-semantics tables to v0.5 — rejected: pattern-parity should land in the milestone that introduces the docs; deferral creates immediate drift (the v0.3 transfer table explicitly verifies within-milestone parity).
 - **Decision:** D-069
 ## Decisions Log
 | ID | Decision | Confidence |
 |----|----------|------------|
 | D-061 | Edge domain owns the proximity/location angle (C4 Locality, C8 Economy heavy); performance/concurrency/observability own generic primitives; cross-links one-directional; edge is a broad-derivation domain (6–8 C-rules) | 0.88 |
 | D-062 | Messaging owns cross-process/cross-system async (brokers, delivery semantics); concurrency owns in-process primitives; boundary is the failure model (network vs crash); cross-link concurrency ← messaging | 0.86 |
 | D-063 | Language-derived docs introduce NO new P-rules; follow existing section-header-trace pattern (C-rule + Domain P-rule inline); first-principles stubs extended but retain structure, no P1–P10 numbering | 0.92 |
 | D-064 | Confirm ts-/py-/go-/rs- prefix scheme for language-derived doc filenames (ecosystem convention alignment; mirrors domain-derived-doc compactness) | 0.84 |
 | D-065 | Add new top-level `## Languages` section to MANIFEST.md in P4 (between Domains and Examples); lists 4 stubs + 16 derived docs; closes pre-existing languages/ drift parallel to ESC-002 | 0.90 |
 | D-066 | principles-matrix.md gains NO language rows (stays 190/19 domains); domain-coverage.md gains a `## Languages Coverage` sub-table (Language → Derived Doc → Traced Domain P-rules → Traced C-rules); inline section-headers are per-doc detail | 0.87 |
 | D-067 | Two phase-specific personas: edge-engineer (P1) + languages-engineer (P3); messaging (P2) by tech-writer + domain-expert with edge-engineer consultation; constraints baked into P2 task must-have (D-051 pattern) | 0.80 |
 | D-068 | Confirm pre-named examples (good: edge-offline-sync, messaging-idempotent-consumer; bad: edge-sync-loop, messaging-shared-subscription); bad examples cite specific breached P-rule; single-breach (not D-044 two-breach) | 0.85 |
 | D-069 | 4 decision/comparison tables across v0.4 derived docs: cdn (edge-cache vs origin), queues (delivery semantics), sync (CRDT vs LWW), streams (Kafka vs Kinesis vs Pulsar), delivery-semantics (dead-letter strategy comparison) — pattern parity with v0.2/v0.3 | 0.82 |
 ## Summary
 - 9 ambiguities identified and auto-resolved at full autonomy
 - Average confidence: 0.86
 - All resolutions grounded in existing project conventions (D-001..D-060, prior domain structure, core C-rules, ARCHITECTURE.md hierarchy, RESEARCH.md prior-art pattern, ESC-002/IDEATE-17 manifest-drift precedent, IDEATE-07/D-025 named-bad-example convention, v0.2→v0.3 decision-matrix transfer pattern, D-051 constraints-baked-into-task persona pattern)
 - No decision fell below the 0.60 autonomy threshold → no HITL escalation triggered
 - Lowest-confidence decision (D-067, 0.80): messaging P2 persona choice — resolved by deferring to the ROADMAP's explicit 2-persona naming and the D-022 within-competence precedent, with edge-engineer consultation to preserve the D-062 boundary
@@ -0,0 +1,193 @@
 # Atelier — Grill (Adversarial Red-Team Review)
 > Pre-execution gate for milestone v0.4 (Edge + Messaging + Language-Derived Docs).
 > Default assumption: the project is unfeasible, over-scoped, and too costly. Not convinced until evidence forces it.
 > Mode: full autonomy. Auto-resolve at confidence ≥ 0.60; escalate only < 0.60 that cannot be auto-resolved.
 ---
 ## v0.4 Grill — 2026-08-05
 **Milestone:** v0.4 — Edge + Messaging + Language-Derived Docs
 **Phase:** 0 (Pre-Execution, GRILL stage)
 **Grill scope:** all 9 axes + meta
 **Prior grill runs:** v0.3 (GRILL.md — PROCEED @ 0.80). v0.1/v0.2 P0 stages did not include a GRILL stage.
 **Autonomy:** full (binding verdicts; no HITL questions)
 **Axes:** 9 review axes (Feasibility, Scope, Cost/Risk, Vertical-slice integrity, Traceability, Decision quality, Persona coverage, Test/Verify, Cross-link/matrix integrity)
 ### Verdict: **CONDITIONAL PROCEED** (overall confidence 0.83)
 The plan is the 4th consecutive NFR milestone using the same proven docs-only contract (v0.1 35 reqs → v0.2 24 reqs → v0.3 32 reqs → v0.4 26 reqs). All three prior milestones shipped clean with identical structure. The 26-req scope is *smaller* than v0.3's 32 reqs, and per-phase doc counts (P1:5, P2:5, P3:20, P4:7, P5:5, P6:2) are within demonstrated throughput. The two new domains (edge, messaging) have well-reasoned boundaries against existing domains (D-061 proximity angle; D-062 cross-process failure model) and the 16-doc language extension is the largest single phase (P3) but introduces **zero** matrix rows (D-063/D-066), so it does not stress the traceability invariant.
 The verdict is **CONDITIONAL** rather than clean PROCEED because the red-team verification surfaced **one false premise** baked into the plan and the audit: IDEATE-34/ATELIER-106 refinement asserts a "pre-existing matrix discrepancy (171 vs stated 170)" that must be reconciled before appending the 20 v0.4 rows. Direct verification of `matrix/principles-matrix.md` shows the matrix contains **exactly 170 P-rows** (17 domains × 10 rows each), matching the invariant — the discrepancy does not exist. The reconciliation task (04-01-01) and audit test (06-01-02) are built on a wrong premise. This is non-blocking (the verification step is harmless — it will simply confirm 170) but the task wording is misleading and could cause an executor to "fix" a non-existent problem (e.g., deleting a valid row to hit a phantom 171→170). Two binding conditions below fix this before ship.
 ---
 ### Axis 1 — Feasibility
 | # | Forcing Question | Evidence | Answer | Confidence |
 |---|-----------------|----------|--------|------------|
 | 1.1 | Can P1–P6 actually be delivered with the proposed resources/personas? | PLAN-v0.4.md: 46 tasks across 6 phases. PERSONAS.md: 3 permanent + 2 phase-specific personas, peak 5 active (matches v0.3). Per-phase doc counts: P1=5, P2=5, P3=20, P4=7, P5=5, P6=2. v0.1 P3 shipped 27 derived docs in one phase (RESEARCH-v0.4 §Language-Derived). | Feasible. v0.1 P3 demonstrated 27 derived docs in a single phase; v0.4 P3's 20 language files (4 stubs + 16 derived) is below that watermark, and language-derived docs have a *simpler* structure (section-header-trace, no P1–P10 derivation) than domain derived docs. The 5-persona peak matches v0.3 which shipped clean. | 0.85 |
 | 1.2 | Are edge/messaging P-rules novel enough to not duplicate existing domains (per D-061/D-062)? | RESEARCH-v0.4 §Edge boundary table: edge P6 (geographic/partition cache invalidation) vs `performance/P5 Caching with Intent` (generic C3,C6) — distinct angle. §Messaging boundary table: messaging P6 (broker-backed backpressure, network failure) vs `concurrency/P9 Bounded Queues` (in-process, OOM failure) — distinct failure model. `concurrency/patterns.md` Pattern 5 explicitly scopes to "producer is blocked or signaled" (in-process); messaging extends to network-partition regime. | Novel. The discriminator is sound: edge owns *where compute runs* (geographic locality, C4); performance owns *generic measurement*; concurrency owns *in-process primitives*; messaging owns *cross-process/network-failure-model*. No verbatim duplication. Each domain owns a distinct angle on shared C-rules — the same pattern that distinguished k8s P8 from iac P3 in v0.2 (D-021 precedent). | 0.86 |
 | 1.3 | Is the 16-doc language-derived extension (P3) achievable in one phase? | RESEARCH-v0.4 §Language-Derived: 16 derived docs follow the existing stub's section-header-trace pattern (confirmed in `languages/typescript.md`: `## Type System (C1 Correctness, Data P7 Type Fidelity)`). PLAN-v0.4 P3 Wave 2: 16 tasks at ≤5 concurrent (5-then-5-then-5-then-1, A-008 @ 0.90). D-063: no P1–P10 derivation (apply, don't derive) — lower per-doc effort than domain derived docs. | Achievable. Language-derived docs are structurally simpler than domain derived docs (no first-principles derivation, no matrix rows, no decision matrices — they apply existing P-rules). v0.1 P3 produced 27 derived docs with heavier structure. The 16-doc load is within demonstrated throughput and below v0.1's high-water mark. | 0.83 |
 **Axis 1 confidence: 0.85.** No challenges.
 ### Axis 2 — Scope
 | # | Forcing Question | Evidence | Answer | Confidence |
 |---|-----------------|----------|--------|------------|
 | 2.1 | Is v0.4 over-scoped (2 new domains + 16 language docs + matrix + review + examples + cross-links)? Should it split into v0.4 + v0.5? What would you cut? | PROJECT.md v0.4 section: 2 domains (10 P-rules each = 20 matrix rows), 16 language-derived docs (0 matrix rows), matrix/review/examples/manifest extensions. REQUIREMENTS ATELIER-92..117 = 26 reqs. v0.3 had 32 reqs (4 domains, 40 matrix rows) and shipped clean. | **Not over-scoped — smaller than v0.3.** 26 reqs vs v0.3's 32. The 16 language-derived docs are the volume driver but carry zero matrix-row cost (D-063/D-066), so the traceability surface grows by only 20 rows (edge + messaging), the smallest matrix extension since v0.2. Cutting to v0.4a (2 domains) + v0.4b (languages) would double release overhead with no load benefit — P3's 16 docs are simpler per-doc than domain docs. The current single-milestone plan is the right call. | 0.85 |
 | 2.2 | What would you cut if forced? | (adversarial) | The simplest 80% version: ship edge + messaging domains + matrix/review (drop the 16 language-derived docs to v0.5). This halves the doc count. **But:** the language-derived extension closes pre-existing MANIFEST drift (languages/ unlisted since v0.1, per D-065/IDEATE-31) and is the v0.3 ROADMAP "Future Milestones" candidate explicitly assigned to v0.4. Deferring it re-opens the 2x-deferral zombie risk flagged in v0.3 G-001. The 16 docs are low-cost (apply, don't derive) — keeping them is correct. | 0.82 |
 | 2.3 | Who owns the requirements, and are they frozen? | REQUIREMENTS.md v0.4 reqs ATELIER-92..117 (all pending). Ideation log IDEATE-31..43: 13 accepted, 0 deferred, 0 rejected. Traceability matrix maps phases→reqs. | Frozen post-ideation. 26 reqs across 6 phases. 13 accepted refinements baked into task must-haves (no moving targets). The ideation stage closed with 0 deferred. | 0.88 |
 | 2.4 | What is explicitly out of scope? | PROJECT.md lines 49–55: tooling/linters, translation, agent adapters, per-domain release artifacts, runtime code. D-020: docs-only. D-063: no new P-rules for languages. | Explicit and enumerated: no runtime code, no tooling, no translation, no new core principles, no new matrix rows for languages, no standalone `.ts`/`.py`/`.go`/`.rs` artifacts (D-020 + ATELIER-110 artifact-types extension). The scope boundary is answerable. | 0.88 |
 | 2.5 | Are there hidden requirements disclosed late? | IDEATE-32 → ATELIER-117 (ARCHITECTURE.md v0.4 update — architecture-tree drift, detected during ideation). Verified: `.ciagent/atelier/ARCHITECTURE.md` component tree stops at v0.3 (lines 14–19 list only through compliance). | ATELIER-117 is a *detected* hidden requirement, surfaced and added during ideation (not late in delivery). The architecture-tree drift is real (ARCHITECTURE.md has no edge/messaging/languages-derived entries). IDEATE-32 caught it before execution — the correct time to surface it. No undisclosed hidden requirements remain. | 0.85 |
 **Axis 2 confidence: 0.86.** No challenges.
 ### Axis 3 — Cost/Risk
 | # | Forcing Question | Evidence | Answer | Confidence |
 |---|-----------------|----------|--------|------------|
 | 3.1 | 46 tasks across 6 phases — what's the riskiest task? | PLAN-v0.4 task inventory: P1 01-01-01 (edge first-principles), P4 04-01-01 (matrix extension), P4 04-01-02 (domain-coverage with Languages sub-table), P3 03-02a..d (16 derived docs). | The riskiest task is **04-01-01 (matrix extension + IDEATE-34 reconciliation)** — but the risk is not the work (appending 20 rows is mechanical); the risk is the **false premise** that the matrix has 171 rows vs stated 170. Direct verification shows the matrix has **exactly 170 rows** (17 × 10). The reconciliation step is built on a wrong premise. **Mitigation:** see binding decision G-014 — the task must verify 170 exists and append 20 to reach 190; it must NOT "fix" a phantom 171→170. | 0.80 |
 | 3.2 | What's the cost of a phase failing mid-way? | NFR milestone, docs-only. All phases produce markdown. No runtime, no infra, no migration. | Cost of a phase failing = re-authoring markdown. P3 (largest) failing mid-way wastes the 16 derived-doc tokens but the work is recoverable (edit + re-run). No cascading infrastructure cost, no data loss, no deployment rollback. Rework is bounded and cheap — the same property that made v0.1/v0.2/v0.3 low-risk. | 0.90 |
 | 3.3 | Predictable cost drivers not in the original budget? | PROJECT.md: no runtime code, no infrastructure, no licensing, no external services. RESEARCH-v0.4 §Risks: 10 risks, all content/traceability (no external cost drivers). | None. Docs-only = no licensing, no infra, no security review fees, no support contracts. Token cost is proportional to markdown volume, scoped by REQ count (26). The only cost driver is markdown authoring, which is bounded. | 0.88 |
 | 3.4 | Single project-killing risk? | RESEARCH-v0.4 §Risks: orphaned P-rules, edge/messaging duplication, language-derived new-P-rule violation, MANIFEST drift, runtime artifacts, persona explosion, matrix row-count, decision-matrix omission. | The single risk that would undermine the framework's core value: **language-derived docs introducing new P-rules** (D-063 violation) or **orphaned from domain P-rules** (D-066 violation). Either breaks the traceable hierarchy (the unique value). **Mitigation:** languages-engineer persona constraint "apply, don't derive" + agent-checklist language-derived scope check (ATELIER-108) + P3 Verify (no new P-rules, every section traces ≥1 domain P-rule) + P4 Languages Coverage sub-table (exactly 16 rows, IDEATE-42) + P6 domain-expert verification. Well-mitigated with multiple layers. | 0.84 |
 **Axis 3 confidence: 0.86.** Challenge: IDEATE-34 false premise (G-014). Auto-resolved with binding condition.
 ### Axis 4 — Vertical-slice integrity
 | # | Forcing Question | Evidence | Answer | Confidence |
 |---|-----------------|----------|--------|------------|
 | 4.1 | Is each phase independently shippable per D-053? | PLAN-v0.4 invariants: P1 ships edge self-consistent (matrix in P4); P2 ships messaging self-consistent; P3 ships 16 language docs self-consistent; P4 closes traceability + manifest + architecture; P5 closes examples + cross-links; P6 ships release. Tag sequence v0.3.1→v0.3.6. | Each phase is shippable. P1/P2 ship domains whose P-rules trace to core internally (matrix extension is traceability closure, not internal-consistency). P3 ships language docs whose section-headers trace to existing domain P-rules (the Languages Coverage sub-table that aggregates these lands in P4 — but each derived doc is self-consistent because it traces to already-shipped domain P-rules). P4 closes manifest + matrix + architecture. P5 closes examples + cross-links. This mirrors v0.3's D-053 pattern that shipped clean. | 0.84 |
 | 4.2 | P1 ships edge domain before matrix lands in P4 — acceptable? | D-053 (v0.3): "matrix rows land in P4 — acceptable because the domain is self-consistent; matrix extension is the traceability closure, not a blocker." v0.2/v0.3 both shipped domain phases before P4 matrix extension. | Acceptable. The edge domain's 10 P-rules each name their C-rule derivation *inline* in `first-principles.md` (per PLAN 01-01-01 must-have). The matrix row is the *aggregate* trace; the inline trace is the *per-rule* detail. A domain with correct inline traces is self-consistent even before the matrix row exists. The matrix in P4 is the verifiable closure, not the source of truth. Precedent: v0.2/v0.3 both did this. | 0.86 |
 | 4.3 | P3 ships language docs that trace to domain P-rules — traceable without the matrix update? | D-063/D-066: language-derived docs use section-header-trace pattern (C-rule + Domain P-rule inline). Languages Coverage sub-table lands in P4. `languages/typescript.md` confirms pattern: `## Type System (C1 Correctness, Data P7 Type Fidelity)`. | Traceable without the matrix update. Each language-derived doc's section headers name the traced domain P-rules inline — the traceability is in the doc itself, not the matrix. The Languages Coverage sub-table (P4) is the *aggregate verifiable view*, parallel to how the domain→core matrix is the aggregate view of inline domain traces. A language-derived doc with correct inline section-header traces is self-consistent even before the sub-table exists. | 0.85 |
 | 4.4 | Edge↔messaging cross-links dangle at P1 tag — acceptable? | PLAN 01-02-03 (edge/iot.md): "cross-link `messaging/queues` ... this link dangles until P2, acceptable per D-053, verified bidirectional in P5 ATELIER-114 per IDEATE-40". A-010 @ 0.86. | Acceptable but a genuine temporary broken-link state. edge/iot.md (P1) cross-links messaging/queues.md (P2) — at the v0.3.1 tag, that link 404s. This is acceptable per D-053 (cross-links close in P5) and is the same pattern as v0.3 intra-milestone links. **Mitigation:** P5 cross-link audit (05-02-01) explicitly verifies edge↔messaging links resolve in BOTH directions before P6 ship (IDEATE-40). The temporary dangle is a documented, time-bounded state, not a silent defect. | 0.82 |
 **Axis 4 confidence: 0.84.** No blocking challenges.
 ### Axis 5 — Traceability
 | # | Forcing Question | Evidence | Answer | Confidence |
 |---|-----------------|----------|--------|------------|
 | 5.1 | Do all 20 proposed new P-rules trace to ≥1 core C-rule without duplicating existing domain P-rules? | RESEARCH-v0.4 §Edge: P1→C4,C1; P2→C1,C5; P3→C8,C1; P4→C1,C5; P5→C1; P6→C1,C3; P7→C1,C5; P8→C4,C6; P9→C1,C8; P10→C7,C5. §Messaging: P1→C1,C2; P2→C1,C2; P3→C1; P4→C1,C2; P5→C1,C5; P6→C1,C8; P7→C4,C6; P8→C5,C7; P9→C5,C6; P10→C7,C1. All 20 map to ≥1 C-rule. Edge derives 7 C-rules; messaging derives 7 C-rules. `core/first-principles.md` confirms C1–C8 definitions. | All 20 trace to ≥1 C-rule. None duplicates an existing domain P-rule verbatim: edge P6 (geographic cache invalidation, C1/C3) ≠ performance P5 (generic caching, C3/C6) — different angle + different C-rule mix. messaging P6 (broker backpressure, C1/C8) ≠ concurrency P9 (in-process bounded queue, C1/C8) — same C-rules but different failure model (network vs OOM), per D-062. The matrix invariant (every P-rule → ≥1 C-rule, no orphans) holds. | 0.87 |
 | 5.2 | RESEARCH-v0.4 against `core/first-principles.md` — derivations valid? | C1 Correctness (non-negotiable), C4 Locality (decisions near consequences), C5 Reversibility, C8 Economy (constrained resources). Edge P1 "Proximity is the Design Driver" → C4 Locality (compute near user/data) — valid geographic-locality derivation. Edge P3 "Resources are Constrained" → C8 Economy — valid. Messaging P7 "Partitioning is Intentional" → C4 Locality (ordering locality) — valid. | Derivations are valid. Edge's C4 emphasis (geographic locality) is a sound reading of "decisions and consequences live near each other" — geographic proximity is a locality concern. Messaging's C4 (partition-key = ordering locality) is sound. No forced/contrived derivations detected. The C1 emphasis (correctness under partition/network-failure) is correct — partition tolerance is a correctness property, not a performance preference (edge P1 explicitly states this). | 0.85 |
 | 5.3 | RESEARCH-v0.4 against existing domains (concurrency/patterns.md for messaging; performance/observability for edge) — no duplication? | `concurrency/patterns.md` Pattern 1 (Message Passing, in-process channels), Pattern 5 (Bounded Queue, in-process "producer is blocked or signaled"). `performance` P5 Caching (generic C3,C6). RESEARCH-v0.4 §Messaging boundary: messaging owns cross-process broker-backed analog with network-failure semantics. §Edge boundary: edge owns geographic/partition-aware cache invalidation. | No duplication. The failure-model discriminator (D-062) is the cleanest boundary: concurrency fails by OOM/crash (in-process); messaging fails by network partition/broker-restart (cross-system). Edge's cache angle (P6, partition-aware invalidation, C1/C3) is distinct from performance's generic cache (P5, C3/C6). Cross-links are one-directional outward (D-026 extended) — no back-link edits to v0.1/v0.2/v0.3 content. | 0.86 |
 | 5.4 | Will the matrix stay at exactly 190 P-rules post-v0.4 (D-059)? | Direct verification: `matrix/principles-matrix.md` currently has exactly **170 P-rows** (17 domains × 10, per awk count). RESEARCH-v0.4 §Matrix: +20 rows (10 edge + 10 messaging). | **CRITICAL FINDING:** The matrix currently has exactly 170 rows, matching the invariant. IDEATE-34/ATELIER-106 refinement claims a "171 vs stated 170 discrepancy" that must be reconciled — **this discrepancy does not exist**. Post-v0.4 will be 170 + 20 = 190, exactly as planned. The reconciliation step (PLAN 04-01-01, 06-01-02) is built on a false premise. See binding decision G-014. | 0.88 |
 **Axis 5 confidence: 0.86.** Challenge: IDEATE-34 false premise (G-014) — auto-resolved with binding condition; does not block but must be corrected in task wording.
 ### Axis 6 — Decision quality
 | # | Forcing Question | Evidence | Answer | Confidence |
 |---|-----------------|----------|--------|------------|
 | 6.1 | Are D-054..D-075 well-grounded? | CLARIFY-v0.4.md D-061..D-069 (avg confidence 0.86, range 0.80–0.92). PLAN-v0.4.md D-070..D-073 (range 0.83–0.88). All above 0.80 except D-067 (0.80, messaging persona). PROJECT.md v0.4 section encodes D-054..D-060 (SPECIFY). | Well-grounded. 15 decisions, average 0.85. All ≥0.80. The lowest is D-067 (0.80 — messaging P2 persona choice) which is explicitly justified by deferring to ROADMAP's 2-persona naming + D-022 within-competence precedent + edge-engineer consultation baked into P2 task must-have (D-051 pattern). No decision is below the 0.80 "well-grounded" bar. | 0.85 |
 | 6.2 | Any low-confidence (<0.80) decisions that should be revisited? | CLARIFY-v0.4.md: D-067 @ 0.80 is the floor. PLAN-v0.4 assumptions A-006..A-010 (0.82–0.95). | None below 0.80. D-067 @ 0.80 is the lowest decision; it is justified by the ROADMAP's explicit 2-persona naming (the authoritative signal) and the D-022 precedent. Revisiting it would require a 3rd phase-specific persona (messaging-engineer), which the RESEARCH explicitly rejected to avoid persona-explosion (roster would hit 6, exceeding v0.3's 5 peak). The constraint-baked-into-task mitigation (PLAN 02-01-01 must-have: "P1/P2 demonstrated edge + messaging boundary (D-062)") preserves the distributed-systems lens without the persona. No revisit needed. | 0.83 |
 | 6.3 | Any decisions contradict prior decisions (D-001..D-053)? | D-053 (vertical-slice integrity) — honored (PLAN invariants line 7). D-020 (docs-only) — honored (PLAN invariants line 13, no `.ts`/`.py`/`.go`/`.rs` artifacts). D-026 (one-directional cross-links) — honored (PLAN 05-02-01 "no back-link edits to v0.1/v0.2/v0.3 content"). D-025 (2+2 examples) — honored (D-068 confirms). | No contradictions. D-053/D-020/D-026/D-025 are all explicitly carried forward. D-063 (no new P-rules for languages) is consistent with ARCHITECTURE.md hierarchy (`languages/` depend on `domains/`). D-066 (no language matrix rows) preserves D-059's 190 total. The v0.4 decisions extend the v0.1→v0.2→v0.3 convention chain without breaking any prior commitment. | 0.88 |
 | 6.4 | Does IDEATE-34 (matrix discrepancy) contradict the actual matrix state? | IDEATE-34 refinement (REQUIREMENTS ATELIER-106 note): "The current matrix has 171 P-rows vs the stated 170 invariant — P4 must reconcile this pre-existing discrepancy." Direct verification: matrix has exactly 170 rows. | **Yes — IDEATE-34 is based on a false premise.** The matrix has 170 rows, not 171. The "reconcile 171→170" instruction in PLAN 04-01-01 and audit 06-01-02 is unfounded. This is a decision-quality defect in the ideation stage: IDEATE-34 (confidence 0.82) asserted a discrepancy without verifying the actual count. **Not a contradiction of a prior decision, but a false factual claim.** See binding decision G-014 — the reconciliation step must be reframed as "verify 170 exists, append 20, confirm 190" — NOT "find and fix a phantom 171st row." | 0.80 |
 **Axis 6 confidence: 0.84.** Challenge: IDEATE-34 false premise (G-014).
 ### Axis 7 — Persona coverage
 | # | Forcing Question | Evidence | Answer | Confidence |
 |---|-----------------|----------|--------|------------|
 | 7.1 | Do edge-engineer + languages-engineer + tech-writer + domain-expert + lead-developer cover all P1–P6 authoring needs? | PERSONAS.md: 3 permanent (lead-developer, tech-writer, domain-expert) + 2 phase-specific (edge-engineer P1, languages-engineer P3). PLAN persona assignments: P1 edge-engineer; P2 tech-writer + domain-expert (edge-engineer consult); P3 languages-engineer; P4 domain-expert/tech-writer/lead-developer; P5 edge-engineer + tech-writer + domain-expert; P6 all 5 review. | Coverage is complete. P1 edge (edge-engineer specialist). P2 messaging (tech-writer + domain-expert generalists with edge-engineer consultation for D-062 boundary — baked into task 02-01-01 must-have per D-051 pattern). P3 languages (languages-engineer specialist). P4 matrix/review/manifest/architecture (domain-expert/tech-writer/lead-developer split by artifact type). P5 examples (edge-engineer for edge examples, tech-writer for messaging). P6 all 5 review then remove 2. No phase lacks an assigned author. | 0.84 |
 | 7.2 | Is messaging P2 under-resourced (no phase-specific persona per D-067)? | D-067 @ 0.80: messaging P2 authored by tech-writer + domain-expert with edge-engineer consultation. CLARIFY-v0.4 §7: rejected messaging-engineer persona (roster would hit 6, exceeding v0.3's 5 peak). PLAN 02-01-01 must-have: "edge-engineer consultation baked in (D-067, D-051 pattern): P1/P2 demonstrated edge + messaging boundary (D-062)." | Under-resourced *relative to P1/P3* (which have specialists) but **adequately resourced** for the work. Messaging's specialty (delivery semantics, idempotency, ordering) is within domain-expert's cross-cutting competence (correctness/observability derivations the domain-expert already validates). The edge-engineer consultation + task must-have preserves the D-062 boundary. RESEARCH-v0.4 §Messaging provides thorough prior-art (Kafka, RabbitMQ, NATS, Jepsen, Reactive Streams, CloudEvents). The risk is content-quality (specialist depth) not traceability (domain-expert validates that). Mitigation is adequate per the v0.3 G-009 precedent (i18n/compliance shipped without specialists). | 0.78 |
 | 7.3 | Is there a key-person dependency? | PERSONAS.md: edge-engineer + languages-engineer are phase-specific, removed post-v0.4. D-051 pattern (constraints baked into task must-haves) survives persona removal. | No key-person risk beyond the v0.3-bounded pattern. Personas are constraint sets, not humans. The languages-engineer's "apply, don't derive" constraint (D-063) is baked into P3 task must-haves AND the agent-checklist (ATELIER-108) AND the P3 Verify block. The edge-engineer's "proximity is correctness" constraint (D-061) is baked into P1 task 01-01-01 must-have. If either persona "fails," the constraints survive in the task specs and review checks. | 0.82 |
 **Axis 7 confidence: 0.81.** Challenge: messaging P2 under-resourcing (G-015) — auto-resolved; bounded by RESEARCH prior-art + domain-expert validation, parallel to v0.3 G-009.
 ### Axis 8 — Test/Verify
 | # | Forcing Question | Evidence | Answer | Confidence |
 |---|-----------------|----------|--------|------------|
 | 8.1 | Are the per-phase Verify blocks in PLAN-v0.4.md sufficient to catch a broken phase before ship? | PLAN-v0.4 each phase has a Verify block: P1 (structural 5 files, behavioral P1–P10 trace, security P9/P6, quality cross-links + decision matrices + D-061 boundary). P2 (similar + edge↔messaging link resolution post-P2). P3 (20 files, no new P-rules, no P1–P10 on stubs, no runtime artifacts, IDEATE-40 cross-links). P4 (matrix 190, coverage sub-table 16 rows, ARCHITECTURE.md component tree, review parity, MANIFEST authoritative). P5 (4 examples, single-breach citation, bidirectional cross-links). P6 (26 reqs covered, reconstruction, matrix 190, MANIFEST, ARCHITECTURE, audit clean). | Sufficient. Each Verify block has structural + behavioral + security + quality checks tied to specific decisions (D-061/D-062/D-063/D-066/D-068/D-069/IDEATE-40/IDEATE-42). The P4 Verify is the strongest: matrix row-count (190), Languages Coverage sub-table (exactly 16 rows — catches orphan language docs), MANIFEST reconstruction (catches unlisted docs), ARCHITECTURE.md component-tree test (catches architecture-tree drift). P6 audit adds the reconstruction + matrix-count + MANIFEST reconstruction triple-test. | 0.86 |
 | 8.2 | What would slip through? | (adversarial) | (a) **Content correctness** — Verify blocks check structure + traceability + cross-links, not technical accuracy of, e.g., the CRDT-vs-LWW decision matrix in sync.md or the Kafka rebalance-strategy enumeration in streams.md. A technically wrong but well-structured doc passes Verify. Mitigation: P6 domain-expert + edge-engineer/languages-engineer review (D-052), but this is one layer, not redundant. (b) **The IDEATE-34 false premise** — the P4/P6 "reconcile 171→170" check is unverifiable against a non-existent discrepancy; an executor could misinterpret it as "delete a row" (harmful) or "confirm 170" (correct). G-014 fixes the wording. (c) **Edge↔messaging dangling links at P1 tag** — structurally acceptable (D-053) but a real broken-link state until P2; only caught bidirectionally in P5, not at the P1/P2 tags. | 0.78 |
 | 8.3 | Is there a working definition of done? | PLAN-v0.4 P6 Verify: 26 reqs covered, reconstruction test passes, matrix exactly 190 (reconciled), MANIFEST reconstruction (incl. Languages section), ARCHITECTURE.md component-tree test, audit clean, tag v0.3.6, branches deleted, personas removed (roster=3). | DoD is concrete and testable: 26 reqs covered, matrix = 190 rows (10 × 19), MANIFEST reconstruction passes (incl. Languages), tag v0.3.6 on main. Not "whatever the demo shows." The matrix-count test (190) and MANIFEST reconstruction are mechanical, automatable checks. | 0.88 |
 **Axis 8 confidence: 0.84.** Challenges: content-correctness gap (inherent to docs projects, bounded by P6 review), IDEATE-34 wording (G-014), dangling links (G-016). All auto-resolved.
 ### Axis 9 — Cross-link / matrix integrity
 | # | Forcing Question | Evidence | Answer | Confidence |
 |---|-----------------|----------|--------|------------|
 | 9.1 | Will the matrix stay at exactly 190 P-rules post-v0.4 (D-059)? | Direct verification: matrix currently has exactly 170 rows. RESEARCH-v0.4 §Matrix: +20 (10 edge + 10 messaging). Languages add 0 (D-063/D-066). | Yes — 170 + 20 = 190. The invariant holds. The IDEATE-34 "171 vs 170" claim is false; see G-014. Post-v0.4 the matrix will have exactly 190 rows (10 per domain × 19), the smallest extension since v0.2. | 0.88 |
 | 9.2 | Will the languages/ sub-table in domain-coverage.md capture all 16 language-derived docs? | ATELIER-107 (IDEATE-42): Languages Coverage sub-table with exactly 16 rows. RESEARCH-v0.4 §Language-Derived sketches the 16-row sub-table (TS quartet + PY quartet + GO quartet + RS quartet). PLAN 04-01-02 must-have: "EXACTLY 16 rows (one per language-derived doc)". P4 Verify: "Languages Coverage sub-table has exactly 16 rows". P6 06-01-01: "domain-expert verifies ... 16 Languages Coverage sub-table rows match the derived docs' actual section-header traces (IDEATE-42)". | Yes — the sub-table is specified at exactly 16 rows, verifiable as a P4 review check, and cross-checked in P6 against the actual section-header traces. The 16 = 4 languages × 4 derived docs. No orphans (every row traced to ≥1 domain P-rule per D-066). The sub-table is the verifiable aggregate; inline section-headers are per-doc detail. | 0.86 |
 | 9.3 | Will the bidirectional edge↔messaging cross-links actually be verified (IDEATE-43/IDEATE-40)? | PLAN 05-02-01 (cross-link audit): "(b) edge↔messaging intra-v0.4 links resolve in BOTH directions — every edge→messaging and messaging→edge link resolves by the P6 ship". P5 Verify: "edge↔messaging intra-v0.4 links resolve bidirectionally (IDEATE-40)". RESEARCH-v0.4 §Review: edge/iot.md → messaging/queues.md (MQTT QoS parallels). | Yes — the P5 cross-link audit (05-02-01) explicitly verifies bidirectional edge↔messaging resolution, AND language-derived docs link to both new (edge/messaging) and existing domains where relevant (e.g., ts-async → messaging/delivery-semantics; go-concurrency → concurrency/patterns + messaging/queues). The audit is a dedicated Wave 2 task after all docs exist, so links are resolvable. The one gap: at the P1 tag (v0.3.1), edge→messaging links dangle (messaging authored in P2) — acceptable per D-053, time-bounded, verified in P5. | 0.84 |
 | 9.4 | MANIFEST integrity — will the Languages section close the pre-existing drift? | MANIFEST.md verified: currently has NO `## Languages` section (sections: Reading Order, Core, Domains, Examples, Cross-Cutting). `languages/{typescript,python,go,rust}.md` exist since v0.1 but are unlisted. D-065/IDEATE-31: add `## Languages` between `## Domains` and `## Examples`. PLAN 04-03-01 must-have: lists 4 stubs + 16 derived docs + Cross-Cutting = "19 domains, 190 P-rules post-v0.4". | Yes — the Languages section is specified to list 4 v0.1 stubs (closing pre-existing drift parallel to ESC-002/ATELIER-91) + 16 new derived docs, positioned per reading order (core→domains→languages→examples). P4 Verify: "MANIFEST is authoritative — every v0.4 file listed, languages/ listed closing pre-existing drift". P6 MANIFEST reconstruction test catches any unlisted doc. | 0.87 |
 **Axis 9 confidence: 0.86.** Challenge: dangling links at P1 tag (G-016) — auto-resolved; time-bounded + P5-verified.
 ### Meta — Closing Review
 | # | Forcing Question | Evidence | Answer | Confidence |
 |---|-----------------|----------|--------|------------|
 | M.1 | If you were the auditor, what would you flag? | (adversarial) | **The IDEATE-34 false premise** — the plan and audit instruct executors to "reconcile a pre-existing 171-vs-170 matrix discrepancy" that does not exist (verified: exactly 170 rows). This is the single most material finding: a verification gate built on a wrong fact. An executor could (a) waste effort hunting a phantom row, or (b) worse, delete a valid row to "fix" 171→170, breaking the invariant. G-014 fixes the wording to "verify 170, append 20, confirm 190." Secondary: content-correctness is one-layer-validated (P6 review only) — inherent to docs projects, bounded. | 0.82 |
 | M.2 | What is the project NOT doing that it should? | PROJECT.md out-of-scope: tooling/linters. | Not doing: automated P-rule→C-rule trace validation (manual domain-expert verification only). At 190 P-rules + 16 language-derived traces, manual verification is approaching the limit where tooling would be warranted — but remains feasible for this milestone. Explicitly out of scope; not a v0.4 blocker. The Languages Coverage sub-table (16 rows) actually *enables* future tooling by making language traces machine-readable. | 0.82 |
 | M.3 | Simplest 80% version? | (adversarial) | Ship edge + messaging domains + matrix/review/manifest/architecture (drop the 16 language-derived docs to v0.5). This cuts P3 entirely (20 files → 0) and reduces the milestone to ~10 reqs. **But:** the language-derived extension closes pre-existing MANIFEST drift (D-065) and is the v0.3 ROADMAP "Future Milestones" candidate explicitly assigned to v0.4. Deferring re-opens the 2x-deferral zombie risk (v0.3 G-001). The 16 docs are low-cost (apply-don't-derive, no matrix rows). Keeping them is the right scope call. | 0.83 |
 | M.4 | What must be true in 90 days for success, and is it true today? | (adversarial) | Must be true: (a) 20 new P-rules trace cleanly to C-rules — RESEARCH-v0.4 pre-maps them, P4 verifies. (b) 16 language-derived docs each trace to ≥1 domain P-rule (no orphans) — P3 Verify + P4 sub-table (16 rows) + P6 cross-check. (c) Content is technically correct (CRDT/LWW, Kafka rebalance, MQTT QoS, delivery semantics) — RESEARCH grounds it, P6 validates. (d) Matrix = exactly 190 — verified today at 170, +20 planned. (a)(b)(d) have explicit verification gates with correct premises (post-G-014). (c) depends on execution quality. All achievable. | 0.83 |
 **Meta confidence: 0.83.**
 ---
 ## Summary Table
 | # | Axis | Verdict | Confidence |
 |---|------|---------|------------|
 | 1 | Feasibility | PASS | 0.85 |
 | 2 | Scope | PASS | 0.86 |
 | 3 | Cost/Risk | CONDITIONAL (G-014: fix IDEATE-34 wording) | 0.86 |
 | 4 | Vertical-slice integrity | PASS | 0.84 |
 | 5 | Traceability | CONDITIONAL (G-014: false premise in reconciliation step) | 0.86 |
 | 6 | Decision quality | CONDITIONAL (G-014: IDEATE-34 factual defect) | 0.84 |
 | 7 | Persona coverage | PASS (G-015 noted, bounded) | 0.81 |
 | 8 | Test/Verify | CONDITIONAL (G-014: audit test premise; G-016: dangling links acknowledged) | 0.84 |
 | 9 | Cross-link / matrix integrity | PASS | 0.86 |
 | — | Meta | PASS | 0.83 |
 | **—** | **Overall** | **CONDITIONAL PROCEED** | **0.83** |
 ---
 ## Binding Decisions
 | ID | Decision | Rationale | Confidence | Alternatives |
 |----|----------|-----------|------------|--------------|
 | G-013 | Proceed with v0.4 scope as planned (2 domains + 16 language-derived docs); do NOT split into v0.4/v0.5 | 26 reqs < v0.3's 32 reqs; language-derived docs carry 0 matrix-row cost (D-063/D-066); v0.1 P3 demonstrated 27 derived docs in one phase; deferral re-opens 2x-zombie risk (v0.3 G-001); the 16 docs are the v0.3 ROADMAP "Future Milestones" candidate explicitly assigned to v0.4 | 0.85 | Split into v0.4a (domains) + v0.4b (languages) — rejected: double release overhead, no load benefit; Drop languages to v0.5 — rejected: 2x deferral zombie risk |
 | G-014 | **CONDITION — fix IDEATE-34 false premise before P4 executes.** The matrix currently has exactly 170 P-rows (verified: 17 domains × 10). There is NO "171 vs 170 discrepancy." PLAN task 04-01-01 and audit 06-01-02 must be reworded from "reconcile pre-existing 171→170 discrepancy" to "verify matrix has exactly 170 rows before appending 20; confirm post-append count = 190." The executor must NOT hunt for or delete a phantom 171st row. | Direct verification of `matrix/principles-matrix.md`: `awk '/^## /{...} /^\| P[0-9]/{count++}'` returns 170 total, 15 lines per section header (10 P-rows + header rows), 17 domain sections. IDEATE-34 (confidence 0.82) asserted the discrepancy without verifying the actual count — a factual defect in the ideation stage. The reconciliation instruction is unfounded and risks an executor breaking the invariant by deleting a valid row. | 0.88 | Leave the wording as-is — rejected: an executor could delete a valid row to "fix" a non-existent discrepancy; Remove the reconciliation step entirely — rejected: keeping a verify-170 step is harmless and good practice, just reworded |
 | G-015 | Proceed with tech-writer + domain-expert for messaging P2 without a messaging-engineer persona (confirms D-067) | Messaging's specialty (delivery semantics, idempotency, ordering) is within domain-expert's cross-cutting competence; edge-engineer consultation + task must-have (02-01-01) preserves D-062 boundary; RESEARCH-v0.4 §Messaging provides thorough prior-art (Kafka/RabbitMQ/NATS/Jepsen/Reactive Streams/CloudEvents); v0.3 G-009 precedent (i18n/compliance shipped without specialists); adding a 3rd phase-specific persona would push roster to 6, exceeding v0.3's 5 peak (persona-explosion risk). Assumption A-011 logged: messaging content correctness depends on RESEARCH prior-art + P6 review. | 0.78 | Add messaging-engineer persona — rejected: persona explosion (roster 6 > v0.3 peak 5); Defer messaging to v0.5 — rejected: scope-creep deferral |
 | G-016 | The edge↔messaging dangling cross-link at P1 tag (edge/iot.md → messaging/queues.md) is acceptable per D-053; the P5 audit (05-02-01) MUST verify bidirectional resolution before P6 ship (confirms IDEATE-40) | D-053 established that cross-links close in P5, not at the domain-phase tag; v0.3 had the same intra-milestone dangling-link pattern; the dangle is time-bounded (resolves at P2 tag v0.3.2); P5 audit explicitly checks both directions (IDEATE-40). Not a silent defect — a documented, verified state. | 0.82 | Block P1 ship until messaging exists — rejected: breaks vertical-slice integrity (D-053), serializes independent work; Remove the edge→messaging link from P1 — rejected: loses the MQTT QoS / delivery-semantics parallel, a high-value cross-link |
 | G-017 | CONFIRM D-063/D-066: language-derived docs introduce NO new P-rules and NO new matrix rows; the Languages Coverage sub-table (exactly 16 rows) is the verifiable aggregate | Direct verification of `languages/typescript.md` confirms the section-header-trace pattern (`## Type System (C1 Correctness, Data P7 Type Fidelity)` — no P1–P10 numbering). ARCHITECTURE.md confirms hierarchy (`languages/` depend on `domains/`). Adding P-rules to languages would break the hierarchy and the 190-row invariant. The 16-row sub-table is the right traceability home (parallel to the per-domain row schema). | 0.87 | Add P1–P10 to languages — rejected: breaks hierarchy + matrix invariant; Add language rows to principles-matrix.md — rejected: languages have no P-rules to map (D-063) |
 | G-018 | CONFIRM D-065/IDEATE-31: MANIFEST gains a `## Languages` section closing pre-existing drift since v0.1 (parallel to ESC-002/ATELIER-91) | MANIFEST.md verified: NO `## Languages` section exists despite `languages/{typescript,python,go,rust}.md` existing since v0.1. This is pre-existing drift on the exact ESC-002 precedent (examples/ was unlisted, closed in v0.3). v0.4 touches `languages/` for the first time since v0.1, so it closes the drift. | 0.87 | Defer stubs to a future milestone, list only 16 new docs — rejected: half-closes the drift, inconsistent with ESC-002 precedent; List languages under `## Domains` — rejected: languages are not domains (ARCHITECTURE.md hierarchy) |
 ### Escalations
 **None.** All challenges auto-resolved at confidence ≥ 0.78 (range: 0.78–0.88). At full autonomy, the IDEATE-34 false premise (G-014) is the lowest-confidence finding but is auto-resolvable by rewording the task (not an escalation — it does not require a human decision, just a factual correction). No axis scored below 0.60 on any forcing question.
 ### Assumptions Logged (this grill)
 | # | Assumption | Confidence |
 |---|-----------|------------|
 | A-011 | Messaging P2 content correctness (delivery-semantics, Kafka rebalance, MQTT QoS) depends on RESEARCH-v0.4 prior-art quality + P6 domain-expert/edge-engineer review, not on a specialist messaging-engineer persona. If P6 surfaces fundamental content errors, the remedy is P2 rework (bounded — markdown edits). Parallel to v0.3 A-006 (i18n/compliance without specialists). | 0.78 |
 ### Conditions for SHIP (CONDITIONAL verdict)
 The phase-0 plan is sound and P1 may proceed immediately. The following condition MUST be addressed before P4 executes (it does not block P1/P2/P3):
 1. **G-014 (mandatory):** Reword the IDEATE-34/ATELIER-106 reconciliation instructions in PLAN-v0.4 task 04-01-01 and audit task 06-01-02. The matrix currently has exactly 170 rows — there is no 171-vs-170 discrepancy. The task must say "verify matrix has exactly 170 rows; append 20 (10 edge + 10 messaging); confirm post-v0.4 count = 190" — NOT "reconcile a pre-existing 171→170 discrepancy." This is a factual correction to prevent an executor from hunting for or deleting a phantom row. The orchestrator should apply this rewording before P4 (or the P4 executor should interpret "reconcile" as "verify 170 exists" rather than "find a missing row").
 No other conditions block ship. The plan is the 4th consecutive NFR milestone using the same proven structure; all prior milestones shipped clean.
 ---
 ## Overall Verdict
 **CONDITIONAL PROCEED** at overall confidence **0.83**.
 The v0.4 phase-0 plan (spec → clarify → research → ideate → plan → grill) is sound: the 26-req scope is smaller than v0.3's 32, the two new domains have well-reasoned boundaries (D-061 proximity angle; D-062 cross-process failure model), the 16-doc language extension carries zero matrix-row cost, all 20 new P-rules trace to ≥1 C-rule, and the verification gates (matrix=190, Languages sub-table=16 rows, MANIFEST reconstruction, ARCHITECTURE.md component-tree test, bidirectional edge↔messaging cross-link audit) are concrete and testable.
 The single CONDITIONAL is the **IDEATE-34 false premise** (G-014): the plan instructs executors to "reconcile a pre-existing 171-vs-170 matrix discrepancy" that direct verification shows does not exist (the matrix has exactly 170 rows). This is a factual defect in the ideation stage, not a structural flaw. It is non-blocking for P1/P2/P3 (which don't touch the matrix count) and is fixed by rewording the P4/P6 task instructions to "verify 170, append 20, confirm 190" — preventing an executor from breaking the invariant by deleting a valid row to fix a phantom discrepancy.
 **P1 may proceed immediately.** The G-014 rewording must land before P4 executes. No escalation to human required (confidence 0.83 ≥ 0.60).
@@ -0,0 +1,154 @@
 # Atelier v0.4 — Ideate Stage
 > Generated during v0.4 P0 IDEATE stage. Autonomy level: `full`
 > (decision_confidence_threshold = 0.60 → auto-accept ≥ 0.6, auto-reject
 > below). Interactive mode used internally but auto-decided; no HITL.
 > Ideas continue from IDEATE-30 (last v0.3 idea). Tier 3 (cross-project)
 > skipped silently — only one project in registry (per workflow error
 > recovery). All ideas are within v0.4 scope (ATELIER-92..116 + one new
 > req ATELIER-117).
 ## Scope
 Refine the v0.4 requirements (ATELIER-92..116) and the RESEARCH-v0.4.md
 proposed content, parallel to the v0.2 (IDEATE-01..16) and v0.3
 (IDEATE-17..30) ideation logs. Sources: git-native pattern mining
 (lessons, decisions D-001..D-072, ESC-001/002, compound solutions),
 coverage-gap analysis (pending v0.4 reqs), verification-layer inversion,
 architectural-drift detection (MANIFEST vs actual files, ARCHITECTURE.md
 vs directory tree, matrix row-count invariant), spec-driven improvement,
 backend-enriched best-practice/chaos ideation for edge/messaging/
 languages.
 ## Counts
 - **Generated:** 13 ideas (mechanical: 8, backend-enriched: 5)
 - **Accepted:** 13 (all v0.4-scope, confidence ≥ 0.79, above 0.6 autonomy threshold → auto-accepted)
 - **Skipped:** 0
 - **New requirements added:** 1 (ATELIER-117 — ARCHITECTURE.md v0.4 update, from IDEATE-32)
 - **Refinements to existing reqs:** 12 (IDEATE-31, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43)
 - **By tier:** mechanical 8, backend-enriched 5, cross-project 0 (skipped silently — single-project registry)
 - **By category:** drift 3, coverage 3, spec 2, quality 2, chaos 1, improvement 2
 ## Ideas Table
 | IDEATE-ID | Source | Category | Confidence | Decision | Mapped REQ |
 |-----------|--------|----------|------------|----------|------------|
 | IDEATE-31 | mechanical (MANIFEST drift, parallel to ESC-002) | drift | 0.86 | accepted → refines | ATELIER-111 (languages/ section must list 4 v0.1 stubs + 16 derived docs) |
 | IDEATE-32 | mechanical (ARCHITECTURE.md vs directory tree) | drift | 0.88 | accepted → new req | ATELIER-117 (update ARCHITECTURE.md for v0.4: component tree + v0.4 section) |
 | IDEATE-33 | mechanical (IDEATE-19 pattern transfer) | coverage | 0.90 | accepted → refines | ATELIER-107 (Core Principle Coverage C1–C8 → domains table update for edge + messaging) |
 | IDEATE-34 | mechanical (matrix row-count invariant) | drift | 0.82 | accepted → refines | ATELIER-106 (verify post-v0.4 = exactly 190; verify current = exactly 170 per grill G-014 — the "171 vs 170" claim was a false premise, corrected) |
 | IDEATE-35 | mechanical (D-069 vs req-row text) | spec | 0.89 | accepted → refines | ATELIER-100, ATELIER-101 (req text must reference streams + dead-letter comparison tables) |
 | IDEATE-36 | mechanical (RESEARCH anti-patterns vs ATELIER-110 text) | quality | 0.85 | accepted → refines | ATELIER-110 (language-derived-doc anti-patterns + .ts/.py/.go/.rs artifact-type extension) |
 | IDEATE-37 | backend-enriched (chaos anti-pattern enumeration) | chaos | 0.86 | accepted → refines | ATELIER-110 (include all 6 RESEARCH chaos anti-patterns: edge sync-loop/TTL-less/blocking-call; messaging shared-sub/blocking-consumer/unbounded-retry) |
 | IDEATE-38 | backend-enriched (IDEATE-28 pattern transfer) | improvement | 0.80 | accepted → refines | ATELIER-94 (offline-first.md maps offline write-queue patterns to testing pyramid tiers) |
 | IDEATE-39 | backend-enriched (IDEATE-29 pattern transfer) | improvement | 0.79 | accepted → refines | ATELIER-101 (delivery-semantics.md fenced idempotency-key dedup-store example, not prose-only) |
 | IDEATE-40 | backend-enriched (intra-milestone cross-link dependency) | chaos | 0.84 | accepted → refines | ATELIER-114 (P5 cross-link audit must verify edge↔messaging intra-v0.4 links resolve in both directions) |
 | IDEATE-41 | backend-enriched (IDEATE-30 pattern transfer) | improvement | 0.81 | accepted → refines | ATELIER-100 (streams.md enumerates consumer-group rebalance strategies: sticky/cooperative/eager) |
 | IDEATE-42 | mechanical (D-066 sub-table verifiability) | coverage | 0.87 | accepted → refines | ATELIER-107 (Languages Coverage sub-table = exactly 16 rows; P4 review check) |
 | IDEATE-43 | mechanical (IDEATE-18 pattern transfer) | coverage | 0.83 | accepted → refines | ATELIER-111 (MANIFEST Cross-Cutting summary = "19 domains, 190 P-rules post-v0.4") |
 ## Refinements Notes (applied to existing reqs at execute time, not changing req rows)
 - **ATELIER-106** (IDEATE-34, corrected by grill G-014): matrix coverage summary must read
  "post-v0.4: 19 domains, 190 P-rules" in BOTH the summary block and
  the per-domain section count (IDEATE-18 carried forward). P4 must
  verify the matrix contains exactly 190 P-rows (10 per domain × 19).
  The current matrix has **exactly 170 P-rows** (17 domains × 10,
  verified by grill G-014 direct count) — the IDEATE-34 "171 vs 170"
  claim was a FALSE PREMISE. P4 must verify 170 exists, append 20 rows,
  and confirm post-v0.4 count = 190. Do NOT hunt for or delete a phantom
  171st row.
 - **ATELIER-107** (IDEATE-33, IDEATE-42): the domain-coverage.md
  extension must update THREE artifacts: (a) the per-domain row-schema
  table with edge + messaging rows; (b) the "Core Principle Coverage"
  table (C1–C8 → domains) for the 2 new domains (C4 Locality grows by
  edge P1/P8 + messaging P7; C5 Reversibility grows by edge P2/P4/P7/P10
  + messaging P5/P8/P9; C7 Observability grows by edge P10 + messaging
  P8/P10; C8 Economy grows by edge P3/P9 + messaging P6; C2 grows by
  messaging P1/P2/P4) — parallel to v0.3 IDEATE-19; (c) the new
  `## Languages Coverage` sub-table with EXACTLY 16 rows (one per
  language-derived doc), verifiable as a P4 review check (no orphans,
  every derived doc traced to ≥1 domain P-rule).
 - **ATELIER-110** (IDEATE-36, IDEATE-37): the anti-patterns extension
  must include (a) the v0.4 deployable-artifact-types table extended to
  `.ts`/`.py`/`.go`/`.rs` standalone files (language-derived docs risk,
  per D-020 docs-only contract); (b) language-derived-doc anti-
  patterns: introducing a new P-rule (e.g., "TS P1"), orphan section
  with no domain P-rule trace, runtime standalone artifact, and
  first-principles stub gaining P1–P10 numbering; (c) all 6 RESEARCH-
  specified chaos anti-patterns, not just the 4 named in the req-row
  text — edge: sync conflict loop (P4), TTL-less cache (P6), blocking
  call on constrained device (P3/P5); messaging: shared subscription
  (P2/P3), blocking consumer (P6), unbounded retry / no DLQ (P5).
 - **ATELIER-100** (IDEATE-35, IDEATE-41): streams.md must include (a)
  the stream-platform comparison table (Kafka vs Kinesis vs
  Pulsar/NATS JetStream — ordering, partitioning, replay/retention,
  use-case fit) per D-069, AND (b) an enumeration of consumer-group
  rebalance strategies (sticky/cooperative vs eager) as a sub-
  comparison — each with partition-stop-the-world cost and use-case
  fit. Parallel to v0.3 IDEATE-30 (drift-type enumeration).
 - **ATELIER-101** (IDEATE-35, IDEATE-39): delivery-semantics.md must
  include (a) the dead-letter strategy comparison table (retry-count-
  limit vs TTL-with-backoff vs poison-queue vs DLQ+alert) per D-069,
  AND (b) a fenced idempotency-key dedup-store example (not prose-
  only) — parallel to v0.3 IDEATE-29 (fenced signed-attestation
  example). Prose-only delivery-semantics guidance is weak; a fenced
  consumer-with-dedup-store demonstrates P3 (idempotent consumers)
  concretely.
 - **ATELIER-111** (IDEATE-31, IDEATE-43): the new `## Languages`
  MANIFEST section must list the 4 v0.1 first-principles stubs
  (typescript.md, python.md, go.md, rust.md — closing the pre-existing
  drift since v0.1, parallel to ESC-002/ATELIER-91) AND the 16 new
  derived docs. The Cross-Cutting summary line must update to exactly
  "19 domains, 190 P-rules post-v0.4" (IDEATE-18 carried forward).
 - **ATELIER-114** (IDEATE-40): the P5 cross-link audit must verify
  edge↔messaging intra-v0.4 links resolve in BOTH directions, not only
  new→existing. Edge/iot.md (P1) cross-links messaging/queues.md (MQTT
  QoS parallels delivery semantics) and messaging/iot.md-style
  references — but messaging/ docs are authored in P2, so at the P1
  tag (v0.3.1) those links dangle. This is acceptable per vertical-
  slice integrity (D-053 — matrix/cross-links close in P4/P5), but the
  P5 audit (ATELIER-114) must explicitly check that every edge→
  messaging and messaging→edge link resolves by the P6 ship. The
  audit must also confirm language-derived docs (P3) link to BOTH
  new (edge/messaging) and existing domains where relevant (e.g.,
  ts-async.md → messaging/delivery-semantics; go-concurrency.md →
  concurrency/patterns + messaging/queues).
 - **ATELIER-94** (IDEATE-38): offline-first.md must map offline write-
  queue / conflict-detection patterns to the testing pyramid tiers
  (unit: conflict detection on a merge function; integration:
  reconnect reconcile against a local store; e2e: partition
  simulation with a fake network) — parallel to v0.3 IDEATE-28 (i18n
  pseudo-locale → pyramid). Avoids generic "test offline" guidance.
 ## Within-Project Pattern Transfer (v0.1 → v0.2 → v0.3 → v0.4)
 | Prior Lesson | v0.4 Application | Status |
 |--------------|------------------|--------|
 | IDEATE-10/11/12 + 21/22/23 (decision/comparison tables in multi-tool derived docs) | D-069 + IDEATE-35 transfer the pattern to streams.md (Kafka/Kinesis/Pulsar) + delivery-semantics.md (dead-letter strategies) | ✓ transferred |
 | IDEATE-17 (manifest drift closure — examples/ unlisted → ATELIER-91) | IDEATE-31 → ATELIER-111 (languages/ unlisted since v0.1 — closes pre-existing drift on the same ESC-002 precedent) | ✓ transferred |
 | IDEATE-18 (coverage summary totals must update — string + per-domain count) | IDEATE-43 → ATELIER-111 (Cross-Cutting = "19 domains, 190 P-rules post-v0.4") + IDEATE-34 → ATELIER-106 (summary block) | ✓ transferred |
 | IDEATE-19 (Core Principle Coverage C1–C8 → domains table update) | IDEATE-33 → ATELIER-107 (C-rule coverage table extended for edge + messaging with the documented C-count shifts) | ✓ transferred |
 | IDEATE-20 (anti-patterns specificity: domain violations + artifact types) | IDEATE-36 + IDEATE-37 → ATELIER-110 (language-derived anti-patterns + .ts/.py/.go/.rs artifact types + all 6 chaos anti-patterns) | ✓ extended |
 | IDEATE-07 (named bad examples cite the P-rule breached) | ATELIER-113 + D-068 already cover (single-breach, cite specific P-rule) | ✓ carried forward |
 | IDEATE-08 (cross-link verification: every new derived doc ≥1 outbound) | ATELIER-114 already covers; IDEATE-40 extends to edge↔messaging intra-milestone links | ✓ extended |
 | IDEATE-13/14 (chaos: orphan P-rule + deployable artifact) | IDEATE-36 → ATELIER-110 (language orphan-section + standalone .ts/.py/.go/.rs runtime artifact) | ✓ extended |
 | IDEATE-28 (pseudo-locale testing → testing pyramid tiers) | IDEATE-38 → ATELIER-94 (offline write-queue testing → pyramid tiers) | ✓ transferred |
 | IDEATE-29 (fenced signed-attestation example, not prose-only) | IDEATE-39 → ATELIER-101 (fenced idempotency-key dedup-store example) | ✓ transferred |
 | IDEATE-30 (drift-type enumeration: data/concept/prediction + signal each) | IDEATE-41 → ATELIER-100 (rebalance-strategy enumeration: sticky/cooperative/eager + cost each) | ✓ transferred |
 ## Summary
 - 13 ideas generated, 13 accepted (confidence ≥ 0.79), 0 skipped.
 - 1 new requirement: ATELIER-117 (ARCHITECTURE.md v0.4 update —
  coverage gap; no existing req covers the architecture-tree drift).
 - 12 refinements baked into ATELIER-94, 100, 101, 106, 107, 110, 111,
  114 — to be honored by the PLAN stage as task must-haves (parallel
  to v0.3's Refinements Notes → Task Bake-In Map).
 - The within-project pattern-transfer convention (decision matrices,
  manifest-drift closure, coverage-summary totals, C-rule coverage
  table, chaos anti-patterns, fenced examples, enumerations, pyramid
  mappings) continues from v0.2→v0.3→v0.4 unchanged.
 - Tier 3 (cross-project) skipped silently: single-project registry.
 - No idea fell below the 0.60 autonomy threshold → no HITL escalation.
@@ -2,7 +2,10 @@
 > Generated during P0 RESEARCH. Atelier is a documentation-only project;
 > the default runtime persona roster is deactivated. Custom personas
-> reflect the docs-as-code nature of the work.
+> reflect the docs-as-code nature of the work. Updated per milestone:
 > v0.4 adds `edge-engineer` (P1) and `languages-engineer` (P3) phase-
 > specific personas (D-067); messaging (P2) is authored by tech-writer
 > + domain-expert with edge-engineer consultation.
 ## Active Personas
@@ -20,7 +23,7 @@
 - **frameworks:** []
 - **constraints:** ["markdown-first", "examples are mandatory", "why over what", "audience awareness"]
 - **territory:** ["**/*.md", "LICENSE", "CHANGELOG.md", "CONTRIBUTING.md"]
- **reason:** Primary author of all framework markdown content. Owns docs-as-code discipline (Documentation P1). Excludes `.ciagent/**` governance files (owned by lead-developer).
+- **reason:** Primary author of all framework markdown content. Owns docs-as-code discipline (Documentation P1). Excludes `.ciagent/**` governance files (owned by lead-developer). For v0.4: co-authors the messaging domain (P2) with domain-expert, per D-067.
 ### domain-expert
 - **active:** true
@@ -28,7 +31,7 @@
 - **frameworks:** []
 - **constraints:** ["every principle traces to core", "10 principles per domain", "no orphans in matrix"]
 - **territory:** ["core/**", "domains/**", "matrix/**", "review/**", "examples/**"]
- **reason:** Validates domain principle correctness and traceability to core principles. Ensures the matrix is complete and no domain rule is orphaned.
+- **reason:** Validates domain principle correctness and traceability to core principles. Ensures the matrix is complete and no domain rule is orphaned. For v0.4: co-authors the messaging domain (P2) with tech-writer, consulting edge-engineer for the D-062 boundary, per D-067.
 ## Inactive Personas (from default roster)
@@ -46,25 +49,45 @@
 ## Phase-Specific Personas
-### platform-engineer (v0.3 — extended, active for v0.3, removed after milestone completion)
+### edge-engineer (v0.4 — REMOVE after milestone completion)
 - **active:** true
 - **phase_specific:** true
 - **domain:** edge computing
 - **frameworks:** []
 - **constraints:** ["latency-bound first", "resource-constrained examples", "trace to core", "10 P-rules per domain", "no runtime code", "proximity is the distinguishing constraint (D-061)", "offline-capability is non-negotiable"]
 - **territory:** ["domains/edge/**", "examples/good/edge-offline-sync.md", "examples/bad/edge-sync-loop.md"]
 - **reason:** Per D-067: edge domain authoring (cdn, offline-first, iot, sync) benefits from a specialist persona with the proximity-is-correctness and offline-is-non-negotiable constraints. The distinguishing angle is proximity/location (D-061) — distinct from performance (generic), concurrency (in-process), and observability (generic telemetry). Active for v0.4 P1 (Edge Domain) only; consulted during P2 (Messaging) for the D-062 in-process-vs-cross-system boundary; removed after v0.4 completes; roster returns to 3 active personas.
 ### languages-engineer (v0.4 — REMOVE after milestone completion)
 - **active:** true
 - **phase_specific:** true
 - **domain:** language-specific application
 - **frameworks:** []
 - **constraints:** ["trace to existing domain P-rules (D-063, D-066) — no new matrix rows", "per-language idioms", "fenced code examples in-language", "10 P-rules per domain does NOT apply to language docs", "manifest gains Languages section (D-065)"]
 - **territory:** ["languages/**"]
 - **reason:** Per D-067: language-derived extension authoring (16 derived docs across TypeScript, Python, Go, Rust) benefits from a specialist persona enforcing the apply-don't-derive discipline (D-063) and the no-new-matrix-rows invariant (D-066). Languages are an application lens, not a principle source — the existing stubs' section-header-trace pattern is preserved, no P1–P10 numbering is introduced. Active for v0.4 P3 (Language-Derived Extension) only; removed after v0.4 completes; roster returns to 3 active personas.
 ### platform-engineer (v0.3 — REMOVED after milestone completion)
 - **active:** false
 - **phase_specific:** true
 - **domain:** infrastructure/platform-automation
 - **frameworks:** []
 - **constraints:** ["declarative-first", "stateless examples", "trace to core", "10 P-rules per domain", "no runtime code", "source-of-truth is git", "reconciliation loop is the primitive"]
 - **territory:** ["domains/gitops-operators/**", "examples/good/gitops-pr.md", "examples/bad/* (gitops-related)"]
- **reason:** Per D-019 / D-027: the v0.2 platform-engineer persona is reused and extended for P1 (gitops-operators), because GitOps/Operators/Progressive Delivery build directly on the k8s + IaC declarative-reconciliation model the persona already embodies. Removed after v0.3 completes; roster returns to 3 active personas. The v0.2 IaC/k8s content remains owned by tech-writer + domain-expert for cross-link maintenance.
+- **reason:** Per D-019 / D-027: the v0.2 platform-engineer persona was reused and extended for P1 (gitops-operators). Removed after v0.3 completed; roster returns to 3 active personas. The v0.3 gitops-operators content remains owned by tech-writer + domain-expert for cross-link maintenance. v0.4: still inactive.
-### ml-engineer (v0.3 — phase-specific, removed after milestone completion)
+### ml-engineer (v0.3 — REMOVED after milestone completion)
- **active:** true
+- **active:** false
 - **phase_specific:** true
 - **domain:** machine-learning engineering
 - **frameworks:** []
 - **constraints:** ["reproducibility is non-negotiable", "data lineage is traceable", "trace to core", "10 P-rules per domain", "no runtime code", "engineering discipline not algorithm design (D-023)", "examples are illustrative markdown only"]
 - **territory:** ["domains/ai-ml/**", "examples/good/ai-ml-reproducibility.md"]
- **reason:** Per D-019 / D-020: AI/ML domain authoring (data versioning, model evaluation, serving, monitoring/drift) benefits from a specialist persona with reproducibility and data-lineage constraints the existing tech-writer persona lacks. Scope is engineering discipline, NOT algorithm/model design (D-023). Removed after v0.3 completes; roster returns to 3 active personas.
+- **reason:** Per D-019 / D-020: AI/ML domain authoring (data versioning, model evaluation, serving, monitoring/drift) benefited from a specialist persona with reproducibility and data-lineage constraints. Scope was engineering discipline, NOT algorithm/model design (D-023). Removed after v0.3 completed; roster returns to 3 active personas. v0.4: still inactive.
 ### platform-engineer (v0.2 — REMOVED after milestone completion)
@@ -74,7 +97,7 @@
 - **frameworks:** []
 - **constraints:** ["declarative-first", "stateless examples", "trace to core", "10 P-rules per domain", "no runtime code"]
 - **territory:** ["domains/infrastructure-as-code/**", "domains/kubernetes/**", "examples/good/terraform-module.md", "examples/good/k8s-deployment.md", "examples/bad/*"]
- **reason:** Specialist authoring for IaC/k8s domain content (terraform, opentofu, state, modules, k8s workloads/networking/storage/rbac/helm/kustomize) where the existing tech-writer persona lacks the domain expertise. Was active for v0.2 P1–P4 only; removed after milestone v0.2 completed (per D-027, D-014). Roster returns to 3 active personas.
+- **reason:** Specialist authoring for IaC/k8s domain content (terraform, opentofu, state, modules, k8s workloads/networking/storage/rbac/helm/kustomize) where the existing tech-writer persona lacks the domain expertise. Was active for v0.2 P1–P4 only; removed after milestone v0.2 completed (per D-027, D-014). Roster returns to 3 active personas. v0.4: still inactive.
 ## Territory Enforcement
@@ -82,9 +105,11 @@ Mode: `warn` (per config.json `personas.territory_enforcement`).
 At `warn`, territory violations are logged but not blocked. This is appropriate for a docs project where tech-writer may touch `.ciagent/` files incidentally (e.g., updating ROADMAP status). Strict mode would be appropriate once territories stabilize.
-## v0.3 Persona Roster Summary
+## v0.4 Persona Roster Summary (active during milestone)
-Active personas for v0.3 (5): lead-developer, tech-writer, domain-expert (span full milestone), platform-engineer (phase-specific, P1 gitops-operators), ml-engineer (phase-specific, P2 ai-ml).
+Active personas (3, span full milestone): lead-developer, tech-writer, domain-expert.
- **Phase assignment:** P1 GitOps/Operators → platform-engineer; P2 AI/ML → ml-engineer; P3 i18n + Compliance → tech-writer + domain-expert (D-022); P4–P5 matrix/review/examples → tech-writer + domain-expert + lead-developer.
+Phase-specific personas (2, active during milestone): edge-engineer (P1 Edge Domain), languages-engineer (P3 Language-Derived Extension) — both removed after v0.4 completion per D-067.
-Inactive personas (3, unchanged): data-engineer, backend-engineer, frontend-engineer.
+Messaging (P2) is authored by tech-writer + domain-expert (the permanent roster) with edge-engineer consultation for the distributed-systems/async overlap (D-062 boundary), per D-067. The edge-engineer's distributed-systems lens survives the P1→P2 persona transition via a task must-have baked into the plan (the D-051 constraints-baked-into-task pattern from v0.3): "P1/P2 demonstrated edge + messaging boundary (D-062); delivery-semantics comparison table required (D-069)."
-Post-v0.3: platform-engineer + ml-engineer removed; roster returns to 3 active personas (lead-developer, tech-writer, domain-expert).
+Inactive default personas (3, unchanged): data-engineer, backend-engineer, frontend-engineer.
 Historical phase-specific personas (3, inactive): platform-engineer (v0.2), platform-engineer (v0.3 extended), ml-engineer (v0.3) — retained as historical record; "v0.4: still inactive" appended to their reasons.
 Peak active roster during v0.4: 5 (3 permanent + 2 phase-specific), matching v0.3's peak and avoiding persona-explosion risk. Roster returns to 3 active personas after v0.4 completion.
@@ -0,0 +1,285 @@
 # Atelier — Plan (v0.4)
 > Vertical-slice plans with wave ordering for milestone v0.4 (Edge + Messaging + Language-Derived Docs). Plans reference REQ-IDs from `.ciagent/atelier/REQUIREMENTS.md` (ATELIER-92..117). NFR milestone — all phases produce docs; no `feat` code. Per `parallelization.max_concurrent_agents = 5`, wave parallelism is capped at 5 concurrent tasks; waves larger than 5 are split into sub-waves. Grounded in `RESEARCH-v0.4.md` (proposed P-rules, derived-doc outlines, anti-patterns, examples, matrix rows) and decisions D-054..D-073.
 ## Plan-Level Invariants
 - **Vertical-slice integrity (D-053 carried forward):** each phase is independently shippable. P1 ships the edge domain self-consistent (matrix rows land in P4 — the domain is internally consistent; matrix extension is traceability closure, not a blocker). P2 ships messaging self-consistent. P3 ships the 16 language-derived docs self-consistent. P4 closes traceability + manifest + architecture. P5 closes examples + cross-links. P6 ships the release. Edge↔messaging intra-milestone cross-links may dangle at the P1 tag (v0.3.1) because messaging is authored in P2; this is acceptable per D-053 — the P5 audit (ATELIER-114) verifies both directions resolve by the P6 ship (IDEATE-40).
 - **Tag sequence:** v0.3.0 (P0) → v0.3.1 (P1) → v0.3.2 (P2) → v0.3.3 (P3) → v0.3.4 (P4) → v0.3.5 (P5) → v0.3.6 (P6 = milestone release). NFR milestone: no separate minor tag; the final patch (v0.3.6) IS the v0.4 deliverable.
 - **Persona lifecycle:** edge-engineer + languages-engineer active P1–P5, review their authored content in P6 Wave 1 (D-052 precedent), removed in P6 Wave 2 (roster returns to 3 active). Messaging (P2) authored by tech-writer + domain-expert with edge-engineer consultation (D-067); the edge-engineer's distributed-systems lens survives the P1→P2 transition via a task must-have baked into P2 tasks (D-051 constraints-baked-into-task pattern). Peak active roster during v0.4: 5 (3 permanent + 2 phase-specific).
 - **Refinements are execute-time must-haves:** all 12 IDEATE-v0.4.md refinements (IDEATE-31, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43) and the refinement notes in REQUIREMENTS.md are baked into the relevant task Must-have columns — they are NOT separate tasks. See the "v0.4 Ideation Refinements → Task Bake-In Map" at the end of this file.
 - **Cross-link rule (IDEATE-08 carried forward):** every new derived doc has ≥1 outbound cross-link to a MANIFEST-listed doc. The P5 audit (ATELIER-114) verifies this AND edge↔messaging intra-v0.4 links resolve in both directions (IDEATE-40) AND language-derived docs link to both new (edge/messaging) and existing domains where relevant.
 - **Anti-pattern pre-specification (RESEARCH-v0.4 §Review Integration):** 6 chaos anti-patterns are pre-specified to reject on sight — edge: sync-conflict-loop (P4), TTL-less-cache (P6), blocking-call-on-constrained-device (P3/P5); messaging: shared-subscription (P2/P3), blocking-consumer (P6), unbounded-retry (P5). These land in `review/anti-patterns.md` in P4; the 2 named bad examples that demonstrate them (edge-sync-loop, messaging-shared-subscription) land in P5.
 - **No runtime code (D-020 carried forward):** all examples and fenced code are illustrative markdown only. The v0.4 deployable-artifact-types table extends to `.ts`/`.py`/`.go`/`.rs` standalone files (forbidden in language-derived docs).
 ## Phase 0 — Pre-Execution (COMPLETE)
 Stages: SPECIFY ✓ → CLARIFY ✓ → RESEARCH ✓ → IDEATE ✓ → PLAN ✓ → GRILL → SHIP
 Branch: `phase/00-pre-execution` (from `milestone/v0.4-edge-quantum-langs`)
 Tag: v0.3.0
 ## Phase 1 — Edge Domain
 **Goal:** Author the `domains/edge/` tree — 10 first principles (P1–P10) plus 4 derived docs (cdn, offline-first, iot, sync). Edge owns the proximity/location angle (D-061): compute near the user/data source, offline as a first-class state, constrained resources declared, sync conflicts bounded, partition-survivable telemetry. Each P-rule derives from core C1–C8 (matrix extension lands in P4).
 **Branch:** `phase/01-edge` (from `milestone/v0.4-edge-quantum-langs`)
 **Personas:** edge-engineer (author), domain-expert (validate traceability), tech-writer (style/format)
 **Tag:** v0.3.1
 **Requirements:** ATELIER-92, ATELIER-93, ATELIER-94, ATELIER-95, ATELIER-96
 ### Wave 1 (sequential — first-principles must exist before derived docs)
 | Task | File | Persona | REQ-ID | Must-have |
 |------|------|---------|--------|-----------|
 | 01-01-01 | `domains/edge/first-principles.md` | edge-engineer | ATELIER-92 | 10 P-rules P1–P10 per RESEARCH-v0.4 §Edge (P1 Proximity is the Design Driver, P2 Offline is a First-Class State, P3 Resources are Constrained and Declared, P4 Sync Conflicts are Bounded Not Infinite, P5 Edge Operations are Idempotent, P6 Cache Invalidation is Explicit, P7 Partial Degradation is Engineered, P8 Geographic Distribution is a First-Class Constraint, P9 Identity is Constrained at the Edge, P10 Edge Observability Survives Partition); each names the core C-rule(s) it derives from; each has definition + "what violates"; broad-derivation domain (7 C-rules: C1, C3, C4, C5, C6, C7, C8); edge-engineer constraints enforced (proximity is the distinguishing constraint per D-061, offline-capability is non-negotiable, latency is a correctness constraint) |
 ### Wave 2 (parallel — 4 derived docs, independent; ≤5 concurrent)
 | Task | File | Persona | REQ-ID | Must-have |
 |------|------|---------|--------|-----------|
 | 01-02-01 | `domains/edge/cdn.md` | edge-engineer | ATELIER-93 | CDN/edge-cache placement + invalidation per RESEARCH-v0.4 §Edge (cache key design, TTL vs explicit invalidation, cache-hit/miss/origin-fetch, origin shielding, purge strategies URL vs soft vs surrogate-key, cache poisoning prevention, multi-CDN routing P8); **decision matrix required (D-069): edge-cache vs origin — when to serve from the pop, when to fetch fresh, when to shield the origin**; cross-link `performance/frontend` (generic caching), `security/input-validation` (cache poisoning), `observability/metrics` (cache-hit ratio); ≥1 outbound cross-link to a MANIFEST-listed doc (IDEATE-08) |
 | 01-02-02 | `domains/edge/offline-first.md` | edge-engineer | ATELIER-94 | Offline as a first-class state (P2), local-first storage (IndexedDB/SQLite/on-device), queue-and-forward for writes, conflict detection on reconnect, UI for offline state, service workers per RESEARCH-v0.4 §Edge; **offline write-queue / conflict-detection patterns mapped to testing pyramid tiers (IDEATE-38, ATELIER-94 refinement): unit (conflict detection on a merge function), integration (reconnect reconcile against a local store), e2e (partition simulation with a fake network)**; fenced code: a service-worker cache strategy + an offline-write-queue sketch; cross-link `concurrency/patterns` (bounded buffer = in-process analog), `observability/logging` (local-first logging); ≥1 outbound cross-link (IDEATE-08) |
 | 01-02-03 | `domains/edge/iot.md` | edge-engineer | ATELIER-95 | Constrained devices at the edge per RESEARCH-v0.4 §Edge: device resource classes (P3), constrained protocols (MQTT, CoAP), device identity + provisioning (P9), telemetry from devices (P10), command idempotency (P5), partial-degradation when devices drop (P7); fenced code: an MQTT publish/subscribe payload with QoS levels + a device-provisioning manifest; cross-link `security/secrets` (device credentials), `messaging/queues` (MQTT QoS parallels delivery semantics — one-directional outward; this link dangles until P2, acceptable per D-053, verified bidirectional in P5 ATELIER-114 per IDEATE-40); ≥1 outbound cross-link (IDEATE-08) |
 | 01-02-04 | `domains/edge/sync.md` | edge-engineer | ATELIER-96 | Reconciling divergent state across partitions per RESEARCH-v0.4 §Edge: the sync problem (P4), conflict resolution strategies, CRDTs (state-based + operation-based), last-write-win (LWW) with vector clocks/timestamps, merge semantics, conflict-free vs conflict-tolerant data types; **decision matrix required (D-069): CRDT vs last-write-win — when each applies, the correctness cost, the operational cost**; fenced code: a CRDT register (LWW-element-set) + an LWW-with-vector-clock merge; cross-link `data/migrations` (schema migration under sync), `concurrency/patterns` (immutability aids merge); ≥1 outbound cross-link (IDEATE-08) |
 **Verify (P1):**
 - Structural: 5 files exist under `domains/edge/`
 - Behavioral: every P1–P10 in first-principles names ≥1 C-rule (domain-expert sign-off); broad-derivation 7 C-rules (C1, C3, C4, C5, C6, C7, C8) confirmed
 - Security: P9 (Identity is Constrained at the Edge) + P6 (Cache Invalidation is Explicit) sections present
 - Quality: each derived doc has ≥1 outbound cross-link to a MANIFEST-listed doc (IDEATE-08); cdn.md carries the edge-cache-vs-origin decision matrix (D-069); sync.md carries the CRDT-vs-LWW decision matrix (D-069); offline-first.md maps offline write-queue to testing pyramid tiers (IDEATE-38); no duplication of performance/concurrency/observability P-rules (D-061 boundary — proximity/location angle only)
 ## Phase 2 — Messaging Domain
 **Goal:** Author the `domains/messaging/` tree — 10 first principles (P1–P10) plus 4 derived docs (queues, pubsub, streams, delivery-semantics). Messaging owns the cross-process / network-failure-model angle (D-062): messages as contracts, explicit delivery semantics, idempotent consumers, bounded backpressure, observable lag. Concurrency owns the in-process analog; cross-link is one-directional outward. Each P-rule derives from core C1–C8 (matrix extension lands in P4).
 **Branch:** `phase/02-messaging` (from `milestone/v0.4-edge-quantum-langs`)
 **Personas:** tech-writer (author), domain-expert (author + validate traceability), edge-engineer (consult — D-067 distributed-systems/async overlap)
 **Tag:** v0.3.2
 **Requirements:** ATELIER-97, ATELIER-98, ATELIER-99, ATELIER-100, ATELIER-101
 ### Wave 1 (sequential — first-principles first)
 | Task | File | Persona | REQ-ID | Must-have |
 |------|------|---------|--------|-----------|
 | 02-01-01 | `domains/messaging/first-principles.md` | tech-writer + domain-expert (edge-engineer consult) | ATELIER-97 | 10 P-rules P1–P10 per RESEARCH-v0.4 §Messaging (P1 Messages are Contracts, P2 Ordering is a Property Not an Assumption, P3 Consumers are Idempotent, P4 Delivery Semantics are Explicit, P5 Dead-Letter Handling is Defined, P6 Backpressure is Bounded, P7 Partitioning is Intentional, P8 Replay and Retention are Configured, P9 Schemas Evolve Compatibly, P10 Messaging is Observable); each names the core C-rule(s) it derives from; each has definition + "what violates"; broad-derivation domain (7 C-rules: C1, C2, C4, C5, C6, C7, C8); **edge-engineer consultation baked in (D-067, D-051 pattern): P1/P2 demonstrated edge + messaging boundary (D-062 — cross-process/network-failure-model, not in-process concurrency); no duplication of concurrency P-rules** |
 ### Wave 2 (parallel — 4 derived docs, independent; ≤5 concurrent)
 | Task | File | Persona | REQ-ID | Must-have |
 |------|------|---------|--------|-----------|
 | 02-02-01 | `domains/messaging/queues.md` | tech-writer + domain-expert | ATELIER-98 | Point-to-point async delivery via queues per RESEARCH-v0.4 §Messaging: producer/consumer model, ack/nack, visibility timeouts, redelivery, FIFO vs standard queues, prefetch + concurrency, long polling; **comparison table required (D-069): at-least-once vs at-most-once vs exactly-once — semantics, latency cost, implementation cost, when each fits**; fenced code: a producer/consumer pair with idempotency key + a redelivery+DLQ flow; cross-link `concurrency/patterns` (in-process bounded queue — one-directional outward), `observability/metrics` (lag); ≥1 outbound cross-link (IDEATE-08) |
 | 02-02-02 | `domains/messaging/pubsub.md` | tech-writer + domain-expert | ATELIER-99 | Fan-out via publish/subscribe per RESEARCH-v0.4 §Messaging: topic/subscription model, fan-out semantics, shared vs independent subscriptions (the `messaging-shared-subscription` chaos anti-pattern lives here — pre-specified in P4 ATELIER-110), filtering, ordering across subscriptions; fenced code: a publish + two independent subscriptions; cross-link `messaging/streams` (pub/sub vs stream — durability boundary), `observability/metrics` (per-subscription lag); ≥1 outbound cross-link (IDEATE-08) |
 | 02-02-03 | `domains/messaging/streams.md` | tech-writer + domain-expert | ATELIER-100 | Durable-log messaging + stream processing per RESEARCH-v0.4 §Messaging: partitioned log model, offsets, consumer groups, replay, retention windows, stream processing (windowing, joins, exactly-once via transactions); **stream-platform comparison table required (D-069): Kafka vs Kinesis vs Pulsar/NATS JetStream — ordering guarantees, partitioning model, replay/retention, use-case fit**; **consumer-group rebalance strategy enumeration (IDEATE-41, ATELIER-100 refinement): sticky/cooperative vs eager — each with partition-stop-the-world cost + use-case fit (parallel to v0.3 IDEATE-30 drift-type enumeration)**; fenced code: a consumer-group reading from offsets + a transactional exactly-once producer; cross-link `messaging/delivery-semantics` (exactly-once via transactions), `data/schema-design` (stream schema), `observability/tracing` (cross-partition traces); ≥1 outbound cross-link (IDEATE-08) |
 | 02-02-04 | `domains/messaging/delivery-semantics.md` | tech-writer + domain-expert | ATELIER-101 | Correctness properties of delivery per RESEARCH-v0.4 §Messaging: the three semantics (P4), idempotency (P3), ordering (P2), dead-letter strategies (P5), retry budgets + backoff, poison messages; **dead-letter strategy comparison table required (D-069): retry-count-limit vs TTL-with-backoff vs poison-queue vs DLQ+alert — when each applies, failure visibility, operational cost**; **fenced idempotency-key dedup-store example (IDEATE-39, ATELIER-101 refinement): not prose-only — a fenced consumer-with-dedup-store demonstrates P3 concretely (parallel to v0.3 IDEATE-29 fenced signed-attestation)**; fenced code: an idempotent consumer with a dedup store + a DLQ routing rule; cross-link `concurrency/patterns` (retry/backoff in-process analog), `errors/patterns` (errors as data for message failures), `observability/metrics` (DLQ depth as an alert); ≥1 outbound cross-link (IDEATE-08) |
 **Verify (P2):**
 - Structural: 5 files exist under `domains/messaging/`
 - Behavioral: every P1–P10 traces to ≥1 C-rule (domain-expert sign-off); broad-derivation 7 C-rules (C1, C2, C4, C5, C6, C7, C8) confirmed
 - Security: P1 (Messages are Contracts) + P5 (Dead-Letter Handling is Defined) sections present
 - Quality: each derived doc ≥1 outbound cross-link (IDEATE-08); queues.md carries the delivery-semantics comparison table (D-069); streams.md carries the stream-platform comparison table (D-069) AND the rebalance-strategy enumeration (IDEATE-41); delivery-semantics.md carries the dead-letter strategy comparison table (D-069) AND a fenced idempotency-key dedup-store example (IDEATE-39); no duplication of concurrency/patterns.md P-rules (D-062 boundary — cross-process/network-failure-model only); edge↔messaging cross-links resolve bidirectionally post-P2 (verified in P5 ATELIER-114 per IDEATE-40)
 ## Phase 3 — Language-Derived Extension
 **Goal:** Extend each `languages/*.md` from a first-principles-only stub to a full derived-doc tree. 4 languages × (extend stub + 4 derived docs) = 16 new derived docs. Language-derived docs introduce NO new P-rules (D-063) and NO new matrix rows (D-066); each section traces to existing domain P-rules via the section-header-trace pattern. The first-principles stub is extended with a pointer to its derived docs but retains its section structure — no P1–P10 numbering (D-063).
 **Branch:** `phase/03-languages-derived` (from `milestone/v0.4-edge-quantum-langs`)
 **Personas:** languages-engineer (author), domain-expert (validate traceability)
 **Tag:** v0.3.3
 **Requirements:** ATELIER-102, ATELIER-103, ATELIER-104, ATELIER-105
 ### Wave 1 (parallel — 4 first-principles stubs extended, independent; ≤5 concurrent → 4 tasks)
 | Task | File | Persona | REQ-ID | Must-have |
 |------|------|---------|--------|-----------|
 | 03-01-01 | `languages/typescript.md` (extend stub) | languages-engineer | ATELIER-102 | Extend existing first-principles stub with a pointer to its 4 derived docs (ts-types, ts-tooling, ts-async, ts-testing); **retain existing section structure — NO P1–P10 numbering added (D-063)**; existing section-header-trace pattern preserved (e.g., `## Type System (C1 Correctness, Data P7 Type Fidelity)`); languages-engineer constraints enforced (apply don't derive, no new P-rules, no new matrix rows per D-063/D-066) |
 | 03-01-02 | `languages/python.md` (extend stub) | languages-engineer | ATELIER-103 | Extend existing first-principles stub with a pointer to its 4 derived docs (py-types, py-tooling, py-async, py-testing); retain existing section structure — NO P1–P10 numbering (D-063); section-header-trace pattern preserved |
 | 03-01-03 | `languages/go.md` (extend stub) | languages-engineer | ATELIER-104 | Extend existing first-principles stub with a pointer to its 4 derived docs (go-types, go-tooling, go-concurrency, go-testing); retain existing section structure — NO P1–P10 numbering (D-063); section-header-trace pattern preserved |
 | 03-01-04 | `languages/rust.md` (extend stub) | languages-engineer | ATELIER-105 | Extend existing first-principles stub with a pointer to its 4 derived docs (rs-ownership, rs-tooling, rs-async, rs-testing); retain existing section structure — NO P1–P10 numbering (D-063); section-header-trace pattern preserved |
 ### Wave 2 (parallel — 16 derived docs, grouped by language; ≤5 concurrent → executor schedules 5-then-5-then-5-then-1)
 > **Parallelism note:** 16 independent tasks all depend only on Wave 1. The executor schedules at most 5 concurrently per `parallelization.max_concurrent_agents`. The 2a/2b/2c/2d labels organize by language, not a hard sequencing barrier — all sub-waves are in the same dependency tier.
 **Wave 2a — TypeScript (4 derived docs, ≤5 concurrent)**
 | Task | File | Persona | REQ-ID | Must-have |
 |------|------|---------|--------|-----------|
 | 03-02a-01 | `languages/ts-types.md` | languages-engineer | ATELIER-102 | TypeScript type system per RESEARCH-v0.4 §Language-Derived: nominal vs structural typing (branded types), generics, narrowing, utility types, discriminated unions; each section header names ≥1 traced domain P-rule AND core C-rule inline (e.g., `## Nominal vs Structural Typing (C1 Correctness, Data P7 Type Fidelity, API P1 Contract Fidelity)`); fenced TS code in-language (illustrative only, no standalone .ts — D-020); traces to Data P7, API P1, Errors P1, Concurrency P1; ≥1 outbound cross-link to a traced domain doc (IDEATE-08; per IDEATE-40 also cross-link new edge/messaging where relevant, e.g., ts-types → messaging/delivery-semantics if schema-fidelity applies) |
 | 03-02a-02 | `languages/ts-tooling.md` | languages-engineer | ATELIER-102 | TypeScript tooling per RESEARCH-v0.4 §Language-Derived: tsc, ESLint, ts-jest, project references, tsconfig discipline; section headers trace to DevOps P2, DevOps P1, Documentation P1; fenced TS code in-language; ≥1 outbound cross-link (IDEATE-08) |
 | 03-02a-03 | `languages/ts-async.md` | languages-engineer | ATELIER-102 | TypeScript async per RESEARCH-v0.4 §Language-Derived: Promises + AbortSignal, async/await, error handling, cancellation; section headers trace to Concurrency P7, P8, P9, Errors P5; fenced TS code in-language; cross-link `concurrency/patterns` + `messaging/delivery-semantics` (per IDEATE-40 — ts-async → messaging/delivery-semantics where async-retry parallels apply); ≥1 outbound cross-link (IDEATE-08) |
 | 03-02a-04 | `languages/ts-testing.md` | languages-engineer | ATELIER-102 | TypeScript testing per RESEARCH-v0.4 §Language-Derived: Vitest/Jest, mock discipline, type-level tests; section headers trace to Testing P1, P2, P3, P9; fenced TS code in-language; cross-link `testing/pyramid` + `testing/fixtures`; ≥1 outbound cross-link (IDEATE-08) |
 **Wave 2b — Python (4 derived docs, ≤5 concurrent; runs in parallel with 2a — total 8 tasks, capped at 5 → executor schedules 5 then 3)**
 | Task | File | Persona | REQ-ID | Must-have |
 |------|------|---------|--------|-----------|
 | 03-02b-01 | `languages/py-types.md` | languages-engineer | ATELIER-103 | Python type system per RESEARCH-v0.4 §Language-Derived: type hints + Pydantic, mypy/pyright, gradual typing; section headers trace to Data P7, Errors P1, API P1; fenced Python code in-language (no standalone .py — D-020); ≥1 outbound cross-link (IDEATE-08) |
 | 03-02b-02 | `languages/py-tooling.md` | languages-engineer | ATELIER-103 | Python tooling per RESEARCH-v0.4 §Language-Derived: ruff, mypy, poetry, uv, virtualenv discipline; section headers trace to DevOps P2, Documentation P1; fenced Python code in-language; ≥1 outbound cross-link (IDEATE-08) |
 | 03-02b-03 | `languages/py-async.md` | languages-engineer | ATELIER-103 | Python async per RESEARCH-v0.4 §Language-Derived: asyncio, anyio, cancellation, structured concurrency; section headers trace to Concurrency P7, P8, P9; fenced Python code in-language; cross-link `concurrency/patterns` + `messaging/queues` (per IDEATE-40 where applicable); ≥1 outbound cross-link (IDEATE-08) |
 | 03-02b-04 | `languages/py-testing.md` | languages-engineer | ATELIER-103 | Python testing per RESEARCH-v0.4 §Language-Derived: pytest, factory_boy, fixture discipline, parametrize; section headers trace to Testing P1, P2, P3; fenced Python code in-language; cross-link `testing/pyramid` + `testing/fixtures`; ≥1 outbound cross-link (IDEATE-08) |
 **Wave 2c — Go (4 derived docs, ≤5 concurrent; same dependency tier as 2a/2b/2d)**
 | Task | File | Persona | REQ-ID | Must-have |
 |------|------|---------|--------|-----------|
 | 03-02c-01 | `languages/go-types.md` | languages-engineer | ATELIER-104 | Go type system per RESEARCH-v0.4 §Language-Derived: named types, generics, interfaces, type assertion discipline; section headers trace to Data P7, Errors P1, API P1; fenced Go code in-language (no standalone .go — D-020); ≥1 outbound cross-link (IDEATE-08) |
 | 03-02c-02 | `languages/go-tooling.md` | languages-engineer | ATELIER-104 | Go tooling per RESEARCH-v0.4 §Language-Derived: go vet, golangci-lint, go test -race, module discipline; section headers trace to DevOps P2, Concurrency P6; fenced Go code in-language; ≥1 outbound cross-link (IDEATE-08) |
 | 03-02c-03 | `languages/go-concurrency.md` | languages-engineer | ATELIER-104 | Go concurrency per RESEARCH-v0.4 §Language-Derived: goroutines, channels, context, select, sync primitives — Go's distinctive strength earns a dedicated concurrency doc (not go-async); section headers trace to Concurrency P1, P3, P5, P7, P8, P9; fenced Go code in-language; cross-link `concurrency/patterns` + `messaging/queues` (per IDEATE-40 — go-concurrency → concurrency/patterns + messaging/queues where channel/broker analogs apply); ≥1 outbound cross-link (IDEATE-08) |
 | 03-02c-04 | `languages/go-testing.md` | languages-engineer | ATELIER-104 | Go testing per RESEARCH-v0.4 §Language-Derived: table-driven tests, t.Parallel, t.Cleanup, race detector; section headers trace to Testing P1, P2, P3, P9; fenced Go code in-language; cross-link `testing/pyramid` + `testing/fixtures`; ≥1 outbound cross-link (IDEATE-08) |
 **Wave 2d — Rust (4 derived docs, ≤5 concurrent; same dependency tier as 2a/2b/2c)**
 | Task | File | Persona | REQ-ID | Must-have |
 |------|------|---------|--------|-----------|
 | 03-02d-01 | `languages/rs-ownership.md` | languages-engineer | ATELIER-105 | Rust ownership per RESEARCH-v0.4 §Language-Derived: Send/Sync, lifetimes, borrowing, ownership transfer — Rust's distinctive strength earns a dedicated ownership doc (not rs-types); section headers trace to Concurrency P1, P3, Data P7; fenced Rust code in-language (no standalone .rs — D-020); ≥1 outbound cross-link (IDEATE-08) |
 | 03-02d-02 | `languages/rs-tooling.md` | languages-engineer | ATELIER-105 | Rust tooling per RESEARCH-v0.4 §Language-Derived: cargo, clippy, fmt, edition discipline; section headers trace to DevOps P2, Documentation P1; fenced Rust code in-language; ≥1 outbound cross-link (IDEATE-08) |
 | 03-02d-03 | `languages/rs-async.md` | languages-engineer | ATELIER-105 | Rust async per RESEARCH-v0.4 §Language-Derived: tokio, async traits, cancellation, pin; section headers trace to Concurrency P5, P7, P8, P9; fenced Rust code in-language; cross-link `concurrency/patterns` + `messaging/delivery-semantics` (per IDEATE-40 where applicable); ≥1 outbound cross-link (IDEATE-08) |
 | 03-02d-04 | `languages/rs-testing.md` | languages-engineer | ATELIER-105 | Rust testing per RESEARCH-v0.4 §Language-Derived: #[test], proptest, property testing, mock discipline; section headers trace to Testing P1, P2, P3, P9; fenced Rust code in-language; cross-link `testing/pyramid` + `testing/fixtures`; ≥1 outbound cross-link (IDEATE-08) |
 > **Parallelism note:** Wave 2a + 2b + 2c + 2d together = 16 independent tasks. The executor schedules at most 5 concurrently per `parallelization.max_concurrent_agents`; the remaining 11 run as slots free. The 2a/2b/2c/2d labels are organizational (by language), not a hard sequencing barrier — all sub-waves are in the same dependency tier (all depend only on Wave 1).
 **Verify (P3):**
 - Structural: 4 stubs extended + 16 new derived docs exist under `languages/` (20 language files total)
 - Behavioral: every derived-doc section header names ≥1 traced domain P-rule AND core C-rule inline (D-063, D-066); no new P-rules introduced; no P1–P10 numbering on language stubs (D-063); domain-expert verifies no orphan sections (every derived doc traces to ≥1 domain P-rule)
 - Security: no standalone `.ts`/`.py`/`.go`/`.rs` runtime artifacts (D-020 docs-only contract — fenced code in `.md` only)
 - Quality: each derived doc ≥1 outbound cross-link to a traced domain doc (IDEATE-08); language-derived docs link to BOTH new (edge/messaging) and existing domains where relevant (IDEATE-40); the 16 derived docs collectively trace to the domains sketched in RESEARCH-v0.4 §Language-Derived (the Languages Coverage sub-table that captures these traces lands in P4 ATELIER-107)
 ## Phase 4 — Matrix + Review + Manifest + Architecture Integration
 **Goal:** Extend the matrix (+20 P-rule → C-rule mappings, 10 edge + 10 messaging; verify pre-existing 170 rows per G-014, append 20, confirm post-v0.4 is exactly 190), domain-coverage (per-domain rows + Core Principle Coverage table for 2 new domains + new Languages Coverage sub-table with exactly 16 rows), ARCHITECTURE.md (v0.4 section + component tree for edge/messaging/languages-derived), review docs (agent + peer-review + anti-patterns with all 6 chaos anti-patterns + language-derived anti-patterns + .ts/.py/.go/.rs artifact types), and the manifest (new `## Languages` section closing pre-existing drift + post-v0.4 totals). Closes the traceability loop and makes the manifest + architecture authoritative for v0.4.
 **Branch:** `phase/04-matrix-review-manifest` (from `milestone/v0.4-edge-quantum-langs`)
 **Personas:** domain-expert (matrix + coverage + anti-patterns), tech-writer (checklists), lead-developer (manifest + ARCHITECTURE.md — authoritative index + architecture tree)
 **Tag:** v0.3.4
 **Requirements:** ATELIER-106, ATELIER-107, ATELIER-117, ATELIER-108, ATELIER-109, ATELIER-110, ATELIER-111
 ### Wave 1 (sequential — matrix is the arbiter, must be authoritative first; then coverage; then ARCHITECTURE.md)
 | Task | File | Persona | REQ-ID | Must-have |
 |------|------|---------|--------|-----------|
 | 04-01-01 | `matrix/principles-matrix.md` (extend) | domain-expert | ATELIER-106 | Add "Edge" + "Messaging" sections, 10 rows each, format matching v0.1/v0.2/v0.3 tables (rows per RESEARCH-v0.4 §Matrix Integration); **review check: row count per new domain = 10, each row ≥1 C-rule** (IDEATE-02 carried forward); **verify pre-existing count first (IDEATE-34 corrected per G-014, ATELIER-106 refinement): the current matrix has exactly 170 P-rows (17 domains × 10, verified by grill G-014) — confirm 170 exists BEFORE appending the 20 v0.4 rows; do NOT hunt for or delete a phantom 171st row; post-v0.4 count must be exactly 190 (= 170 + 20)**; update Coverage Summary to "post-v0.4: 19 domains, 190 P-rules" in BOTH the summary block AND the per-domain section count (IDEATE-18, IDEATE-43 carried forward) |
 | 04-01-02 | `matrix/domain-coverage.md` (extend) | domain-expert | ATELIER-107 | THREE artifacts updated (IDEATE-33, IDEATE-42, ATELIER-107 refinement): (a) per-domain row-schema table — add Edge + Messaging rows (schema: domain, P-count, derived-doc-count, manifest-listed, status per IDEATE-03); (b) Core Principle Coverage table (C1–C8 → domains) — extend for 2 new domains: C4 Locality grows by edge P1/P8 + messaging P7; C5 Reversibility grows by edge P2/P4/P7/P10 + messaging P5/P8/P9; C7 Observability grows by edge P10 + messaging P8/P10; C8 Economy grows by edge P3/P9 + messaging P6; C2 grows by messaging P1/P2/P4 (parallel to v0.3 IDEATE-19); (c) new `## Languages Coverage` sub-table with EXACTLY 16 rows (one per language-derived doc) — schema: Language \| Derived Doc \| Traced Domain P-rules \| Traced Core C-rules; verifiable as a P4 review check (no orphans, every derived doc traced to ≥1 domain P-rule, per D-066) |
 | 04-01-03 | `ARCHITECTURE.md` (extend) | lead-developer | ATELIER-117 | Extend the component tree with `edge/` + `messaging/` domain entries; add a `## v0.4 Domain Additions` section documenting the two new domains (boundary per D-061/D-062, derived docs, cross-link targets); note the `languages/` derived-doc extension (D-063 apply-don't-derive — no new P-rules, no new matrix rows). Closes architecture-tree drift: the file currently stops at v0.3 with no edge/messaging/languages-derived coverage (IDEATE-32) |
 ### Wave 2 (parallel — 3 review docs, independent; ≤5 concurrent)
 | Task | File | Persona | REQ-ID | Must-have |
 |------|------|---------|--------|-----------|
 | 04-02-01 | `review/agent-checklist.md` (extend) | tech-writer | ATELIER-108 | Add 3 trigger sections per RESEARCH-v0.4 §Review Integration: "If Edge (see domains/edge/)" — 10 checks (P1 proximity/latency-as-correctness, P2 offline, P3 constrained resources declared, P4 sync conflicts bounded, P5 idempotent, P6 cache invalidation explicit, P7 partial degradation, P8 geographic distribution, P9 identity constrained, P10 local-first telemetry); "If Messaging (see domains/messaging/)" — 10 checks (P1 messages are contracts, P2 ordering explicit, P3 idempotent consumers, P4 delivery semantics explicit, P5 dead-letter defined, P6 backpressure bounded, P7 partitioning intentional, P8 replay/retention configured, P9 schemas evolve compatibly, P10 observable); "If Language-Derived Docs (see languages/)" — 5 checks (no new P-rules per D-063/D-066; every section header names ≥1 traced domain P-rule AND C-rule inline; fenced code in-language + illustrative only per D-020; first-principles stub retains structure — no P1–P10 numbering; cross-links to traced domain docs present ≥1 outbound per derived doc per ATELIER-114) |
 | 04-02-02 | `review/peer-review-checklist.md` (extend) | tech-writer | ATELIER-109 | Add 3 peer-review sections mirroring the agent-checklist as reviewer-voice questions, plus 3 scope checks (IDEATE-09 parity carried forward): edge scope check (proximity/location/constraint angle, not generic perf or in-process concurrency — D-061); messaging scope check (cross-process/network-failure-model, not in-process concurrency — D-062); language-derived scope check (apply don't derive — no new P-rules, no new matrix rows — D-063, D-066) |
 | 04-02-03 | `review/anti-patterns.md` (extend) | domain-expert | ATELIER-110 | Per RESEARCH-v0.4 §Review Integration + IDEATE-36/37 (ATELIER-110 refinement): (a) **v0.4 deployable-artifact-types table extended to `.ts`/`.py`/`.go`/`.rs` standalone files** (language-derived docs risk, per D-020 docs-only contract); (b) **all 6 RESEARCH-specified chaos anti-patterns** (not just the 4 named in the req-row text) — edge: sync conflict loop (P4, C1/C5 — infinite oscillation is a correctness failure/livelock), TTL-less cache (P6, C1/C3 — stale-forever under partition), blocking call on constrained device (P3/P5, C8/C1 — blocks the node, no timeout, retry unsafe without idempotency); messaging: shared subscription (P2/P3, C1 — breaks ordering + dedup), blocking consumer (P6, C1/C8 — blocks the partition, lag grows), unbounded retry / no DLQ (P5, C1/C5 — infinite retry is a livelock); (c) **language-derived-doc anti-patterns**: introducing a new P-rule (e.g., "TS P1" — violates D-063 + ARCHITECTURE.md hierarchy), orphan section with no domain P-rule trace (violates D-066 + matrix completeness), runtime standalone artifact `.ts`/`.py`/`.go`/`.rs` (violates D-020), first-principles stub gaining P1–P10 numbering (violates D-063); (d) **domain-specific anti-patterns** per RESEARCH-v0.4 §Review Integration — edge 10 (central-region-only deployment, crash-on-disconnect, undeclared resource budget, sync oscillation, non-idempotent edge write, TTL-less cache, crash-on-node-failure, random geographic placement, shared edge-device credential, fire-and-forget telemetry), messaging 10 (schemaless message, undocumented "FIFO", non-idempotent consumer, unstated delivery semantic, no DLQ, unbounded retry budget, default partition key, ephemeral broker, silent schema change, silent backlog) |
 ### Wave 3 (sequential — MANIFEST after review docs settled; per D-050 precedent anti-patterns + manifest combine under lead-developer so manifest lists file paths after anti-pattern content is settled)
 | Task | File | Persona | REQ-ID | Must-have |
 |------|------|---------|--------|-----------|
 | 04-03-01 | `MANIFEST.md` (extend) | lead-developer | ATELIER-111 | Per D-065 + IDEATE-31/43 (ATELIER-111 refinement): (a) add 2 new domains (edge, messaging) + all 10 derived docs to the Domains table; (b) add a new top-level `## Languages` section between `## Domains` and `## Examples` (matching reading order: core → domains → languages → examples) listing the 4 v0.1 first-principles stubs (typescript.md, python.md, go.md, rust.md — closes pre-existing drift since v0.1, parallel to ESC-002/ATELIER-91) AND the 16 new v0.4 derived docs; section note: "Language docs apply (not derive) domain principles. They introduce no new P-rules and no new matrix rows (D-063, D-066)."; (c) update Cross-Cutting summary line to exactly "19 domains, 190 P-rules post-v0.4" (IDEATE-18, IDEATE-43 carried forward) |
 **Verify (P4):**
 - Structural: matrix has 19 domain sections (17 v0.1/v0.2/v0.3 + 2 new), exactly 190 P-rules (170 pre-existing verified per G-014 + 20 new; NOT 171 reconciled to 170 — the 171 discrepancy was a false premise corrected by grill G-014); domain-coverage has the per-domain v0.4 table AND the updated C-rule coverage table AND the Languages Coverage sub-table (exactly 16 rows); ARCHITECTURE.md has the v0.4 section + component tree; 3 review docs extended; MANIFEST lists all v0.4 docs + the new Languages section
 - Behavioral: every new edge + messaging P-rule has a matrix row; domain-expert verifies no orphans (IDEATE-13 carried forward); Coverage Summary reads "19 domains, 190 P-rules" (IDEATE-18/43); Languages Coverage sub-table has exactly 16 rows — every language-derived doc traced to ≥1 domain P-rule (IDEATE-42, D-066); ARCHITECTURE.md component tree matches the directory tree (IDEATE-32)
 - Security: anti-patterns cover all 6 chaos anti-patterns (edge sync-loop/TTL-less/blocking-call; messaging shared-sub/blocking-consumer/unbounded-retry) + edge P9 (shared credential) + messaging P1 (schemaless) + language-derived runtime-artifact (.ts/.py/.go/.rs)
 - Quality: MANIFEST is authoritative — every v0.4 file listed, languages/ listed closing pre-existing drift (ATELIER-111 per IDEATE-31); unlisted = not part of framework; agent-checklist + peer-review-checklist have parity across edge + messaging + language-derived (ATELIER-108 ↔ ATELIER-109); matrix row-count test passes (exactly 190, reconciled)
 ## Phase 5 — Examples + Cross-Links
 **Goal:** Add 2 good + 2 bad examples (D-068 — pre-named, highest illustration value) and verify cross-domain links from the 2 new domains + 16 language-derived docs to existing ones, AND edge↔messaging intra-v0.4 links resolve in both directions (IDEATE-40). Examples are markdown with fenced code only (no standalone .ts/.py/.go/.rs artifacts — D-020). Bad examples are single-breach citing the specific breached P-rule (D-068).
 **Branch:** `phase/05-examples-crosslinks` (from `milestone/v0.4-edge-quantum-langs`)
 **Personas:** edge-engineer (edge examples), tech-writer (messaging examples + cross-link audit), domain-expert (P-rule citation + cross-link validation)
 **Tag:** v0.3.5
 **Requirements:** ATELIER-112, ATELIER-113, ATELIER-114
 ### Wave 1 (parallel — 4 examples, independent; ≤5 concurrent)
 | Task | File | Persona | REQ-ID | Must-have |
 |------|------|---------|--------|-----------|
 | 05-01-01 | `examples/good/edge-offline-sync.md` | edge-engineer | ATELIER-112 | Good edge example per RESEARCH-v0.4 §Examples: demonstrates offline-first + sync conflict resolution via CRDT or LWW — highest-illustration-value edge case; markdown with fenced code (a CRDT register merge + an offline write queue + a reconnect reconcile); P-rules demonstrated: edge P2 (Offline is a First-Class State), P4 (Sync Conflicts are Bounded), P5 (Idempotent), P10 (local-first telemetry); exercises the `sync.md` CRDT-vs-LWW decision matrix; cross-link to `edge/offline-first.md` + `edge/sync.md`; no runtime artifacts |
 | 05-01-02 | `examples/good/messaging-idempotent-consumer.md` | tech-writer | ATELIER-112 | Good messaging example per RESEARCH-v0.4 §Examples: demonstrates exactly-once-via-idempotency-keys + dedup — highest-illustration-value messaging case; markdown with fenced code (an idempotent consumer with a dedup store + a DLQ routing rule); P-rules demonstrated: messaging P3 (Consumers are Idempotent), P4 (Delivery Semantics Explicit), P5 (Dead-Letter Handling), P10 (Observable); exercises the `delivery-semantics.md` idempotency + DLQ guidance; cross-link to `messaging/delivery-semantics.md` + `messaging/queues.md`; no runtime artifacts |
 | 05-01-03 | `examples/bad/edge-sync-loop.md` | edge-engineer | ATELIER-113 | Bad edge example per RESEARCH-v0.4 §Examples + ATELIER-110: a sync conflict that oscillates/loops forever — the pre-specified edge chaos anti-pattern; **single-breach (D-068): cites edge P4 (Sync Conflicts are Bounded, Not Infinite) breached** — prose explains why it is a livelock, not eventual consistency; markdown with fenced code showing the oscillating sync; cross-link to `edge/sync.md` + `review/anti-patterns.md` (the sync-conflict-loop chaos anti-pattern) |
 | 05-01-04 | `examples/bad/messaging-shared-subscription.md` | tech-writer | ATELIER-113 | Bad messaging example per RESEARCH-v0.4 §Examples + ATELIER-110: a shared subscription causing duplicate processing / fan-out duplication — the pre-specified messaging chaos anti-pattern; **single-breach (D-068): primary breach cited is messaging P2 (Ordering is a Property, Not an Assumption); P3 (Consumers are Idempotent) noted as the compounding consequence — the example remains single-breach in its named violation**; markdown with fenced code showing the shared subscription; cross-link to `messaging/pubsub.md` + `review/anti-patterns.md` (the shared-subscription chaos anti-pattern) |
 ### Wave 2 (sequential — cross-link audit after all docs exist)
 | Task | File | Persona | REQ-ID | Must-have |
 |------|------|---------|--------|-----------|
 | 05-02-01 | Cross-link audit (all new derived docs across edge + messaging + languages) | tech-writer | ATELIER-114 | Per IDEATE-08 (carried forward) + IDEATE-40 (ATELIER-114 refinement): (a) every new derived doc (10 domain derived + 16 language-derived = 26) has ≥1 outbound cross-link to a MANIFEST-listed doc; links resolve; (b) **edge↔messaging intra-v0.4 links resolve in BOTH directions** — every edge→messaging and messaging→edge link resolves by the P6 ship (edge/iot.md → messaging/queues.md MQTT QoS parallels; messaging docs that reference edge where relevant); (c) **language-derived docs link to BOTH new (edge/messaging) and existing domains where relevant** (e.g., ts-async.md → messaging/delivery-semantics; go-concurrency.md → concurrency/patterns + messaging/queues); domain-expert validates the cross-link targets are correct (not just present); no back-link edits to v0.1/v0.2/v0.3 content (D-026 extended — one-directional outward) |
 **Verify (P5):**
 - Structural: 4 new example files exist (all .md)
 - Behavioral: each bad example cites the P-rule breached (edge: P4 single-breach; messaging: P2 primary + P3 consequence, single-breach named violation per D-068); each good example cites the P-rules it demonstrates
 - Security: no standalone `.ts`/`.py`/`.go`/`.rs` runtime artifacts (deployable artifact mitigation, IDEATE-36, D-020)
 - Quality: all cross-links from the 26 new derived docs resolve to MANIFEST-listed docs; edge↔messaging intra-v0.4 links resolve bidirectionally (IDEATE-40); language-derived docs link to both new + existing domains where relevant (IDEATE-40); no back-link edits to v0.1/v0.2/v0.3 content (D-026 extended)
 ## Phase 6 — Final Review + Ship (N+1)
 **Goal:** Multi-persona review across all v0.4 phases, audit, milestone ship. P6 IS the v0.4 release (NFR → no separate minor tag; v0.3.6 IS the deliverable). Phase-specific personas (edge-engineer, languages-engineer) review their authored content in Wave 1 (D-052 precedent), then are removed in Wave 2 (roster returns to 3 active).
 **Branch:** `phase/06-final-review-ship` (from `milestone/v0.4-edge-quantum-langs`)
 **Personas:** lead-developer (coordinate + ship), domain-expert (review), tech-writer (review), edge-engineer (review edge content, then removed), languages-engineer (review language-derived content, then removed)
 **Tag:** v0.3.6 (IS the v0.4 milestone release — NFR, no separate minor tag)
 **Requirements:** ATELIER-115, ATELIER-116
 ### Wave 1 — Review (sequential — edge-engineer + languages-engineer review their content per D-052 before removal)
 | Task | Activity | Persona | REQ-ID | Must-have |
 |------|----------|---------|--------|-----------|
 | 06-01-01 | `ciagent-review` — multi-persona review of all v0.4 changes | lead-developer | ATELIER-115 | Auto-apply P0 fixes; flag P1+ for post-hoc; if P1+ found, fix in this phase; **edge-engineer reviews edge domain content (D-061 boundary — proximity/location angle, not generic perf/concurrency)**; **languages-engineer reviews language-derived content (D-063 apply-don't-derive scope check — no new P-rules, no P1–P10 numbering, every section traces to a domain P-rule)**; tech-writer + domain-expert review messaging content (D-062 boundary — cross-process/network-failure-model); domain-expert verifies all 20 new P-rules (10 edge + 10 messaging) trace to ≥1 C-rule (no orphans) AND the 16 Languages Coverage sub-table rows match the derived docs' actual section-header traces (IDEATE-42) |
 | 06-01-02 | `ciagent-audit` — reconstruction + discipline | lead-developer | ATELIER-115 | git log matches .ciagent/ files; branch hygiene; commit discipline (every commit has `---ci---` block with `project: atelier`); MANIFEST reconstruction test (every listed doc exists, every existing doc is listed — incl. the new Languages section per ATELIER-111); **matrix row-count test: exactly 190 P-rows (10 per domain × 19 — the pre-existing count was exactly 170 per grill G-014 verification; the IDEATE-34 "171-vs-170 discrepancy" was a false premise, corrected — verify 170 + 20 = 190, NOT 171 + 20)**; ARCHITECTURE.md component-tree test (matches directory tree per ATELIER-117/IDEATE-32) |
 ### Wave 2 — Ship (sequential — after review; phase-specific personas removed)
 | Task | Activity | Persona | REQ-ID | Must-have |
 |------|----------|---------|--------|-----------|
 | 06-02-01 | `ciagent-ship` — milestone ship | lead-developer | ATELIER-116 | Merge `phase/06` → `milestone/v0.4-edge-quantum-langs` → `main`; tag `v0.3.6`; Gitea release with full milestone summary; delete all v0.4 branches (tags preserve history) |
 | 06-02-02 | Complete milestone (REQUIREMENTS + ROADMAP + PERSONAS) | lead-developer | ATELIER-116 | Mark all v0.4 requirements (ATELIER-92..117) `covered`; ROADMAP v0.4 → complete; **PERSONAS: remove edge-engineer + languages-engineer (roster returns to 3 active per D-067)**; clear checkpoint |
 **Verify (P6):**
 - Structural: all 26 v0.4 requirements (ATELIER-92..117) marked covered
 - Behavioral: reconstruction test passes (git log ↔ .ciagent/); matrix row-count test passes (exactly 190 — reconciled per IDEATE-34); MANIFEST reconstruction test passes (incl. Languages section); ARCHITECTURE.md component-tree test passes (per ATELIER-117/IDEATE-32)
 - Security: audit clean (no critical issues); no standalone runtime artifacts (`.ts`/`.py`/`.go`/`.rs` forbidden per D-020); no new P-rules in language-derived docs (D-063); edge/messaging boundaries respected (D-061/D-062)
 - Quality: milestone merged to main, tag v0.3.6 exists, all v0.4 branches deleted; edge-engineer + languages-engineer personas removed (roster = 3 active per D-067)
 ## v0.4 Wave Ordering Summary
 | Phase | Waves | Parallelism |
 |-------|-------|-------------|
 | P0 | (pre-exec) | Sequential stages (specify→clarify→research→ideate→plan→grill) — COMPLETE |
 | P1 | 2 | Wave 1 sequential (first-principles), Wave 2 parallel (4 derived docs) |
 | P2 | 2 | Wave 1 sequential (first-principles), Wave 2 parallel (4 derived docs) |
 | P3 | 2 | Wave 1 parallel (4 stubs extended), Wave 2 parallel (16 derived docs — 2a TS + 2b Python + 2c Go + 2d Rust, capped at 5 concurrent) |
 | P4 | 3 | Wave 1 sequential (matrix → coverage → ARCHITECTURE.md), Wave 2 parallel (3 review docs), Wave 3 sequential (MANIFEST after review docs settled) |
 | P5 | 2 | Wave 1 parallel (4 examples), Wave 2 sequential (cross-link audit) |
 | P6 | 2 | Wave 1 sequential (review → audit), Wave 2 sequential (ship → complete) |
 ## v0.4 Requirements → Phase Mapping
 | Phase | Requirements | Count |
 |-------|-------------|-------|
 | 1 | ATELIER-92..96 | 5 |
 | 2 | ATELIER-97..101 | 5 |
 | 3 | ATELIER-102..105 | 4 |
 | 4 | ATELIER-106, 107, 117, 108, 109, 110, 111 | 7 |
 | 5 | ATELIER-112, 113, 114 | 3 |
 | 6 | ATELIER-115, 116 | 2 |
 | **Total** | | **26** |
 ## v0.4 Ideation Refinements → Task Bake-In Map
 All 12 accepted ideation refinements (IDEATE-31, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43) + the 1 new requirement (IDEATE-32 → ATELIER-117) are baked into the relevant phase tasks as explicit must-have notes (parallel to v0.3's Refinements Notes → Task Bake-In Map):
 | IDEATE-ID | Refinement | Baked Into Task(s) | How |
 |-----------|-----------|-------------------|-----|
 | IDEATE-31 | MANIFEST Languages section lists 4 v0.1 stubs + 16 derived docs (refines ATELIER-111) | 04-03-01 | MANIFEST gains `## Languages` section between `## Domains` and `## Examples`; lists 4 stubs (closes pre-existing drift since v0.1, parallel to ESC-002/ATELIER-91) + 16 derived docs |
 | IDEATE-32 | ARCHITECTURE.md v0.4 update (new req ATELIER-117) | 04-01-03 | ARCHITECTURE.md gains `## v0.4 Domain Additions` section + component tree entries for edge/messaging/languages-derived; closes architecture-tree drift |
 | IDEATE-33 | Core Principle Coverage C1–C8 → domains table update for edge + messaging (refines ATELIER-107) | 04-01-02 | domain-coverage.md C-rule coverage table extended (C4 +edge P1/P8 +messaging P7; C5 +edge P2/P4/P7/P10 +messaging P5/P8/P9; C7 +edge P10 +messaging P8/P10; C8 +edge P3/P9 +messaging P6; C2 +messaging P1/P2/P4) |
 | IDEATE-34 | Matrix post-v0.4 = exactly 190; verify current 170 (corrected per G-014 — the "171 vs 170" claim was a false premise; matrix has exactly 170) (refines ATELIER-106) | 04-01-01, 06-01-02 | P4 verifies 170 exists before appending 20 rows; P6 audit verifies exactly 190 (10 × 19). Do NOT hunt for/delete a phantom 171st row |
 | IDEATE-35 | streams.md + delivery-semantics.md comparison tables required (refines ATELIER-100, 101) | 02-02-03, 02-02-04 | streams.md stream-platform comparison (Kafka/Kinesis/Pulsar); delivery-semantics.md dead-letter strategy comparison |
 | IDEATE-36 | Language-derived-doc anti-patterns + .ts/.py/.go/.rs artifact-type extension (refines ATELIER-110) | 04-02-03 | anti-patterns.md gains language-derived anti-patterns (new P-rule, orphan section, runtime artifact, P1–P10 numbering) + v0.4 artifact types table extended to .ts/.py/.go/.rs |
 | IDEATE-37 | All 6 RESEARCH chaos anti-patterns (refines ATELIER-110) | 04-02-03 | anti-patterns.md includes all 6: edge sync-loop/TTL-less/blocking-call; messaging shared-sub/blocking-consumer/unbounded-retry (not just the 4 named in req-row text) |
 | IDEATE-38 | offline-first.md maps offline write-queue to testing pyramid tiers (refines ATELIER-94) | 01-02-02 | offline-first.md maps unit (conflict detection on merge function), integration (reconnect reconcile vs local store), e2e (partition simulation with fake network) |
 | IDEATE-39 | delivery-semantics.md fenced idempotency-key dedup-store example (refines ATELIER-101) | 02-02-04 | delivery-semantics.md includes a fenced consumer-with-dedup-store (not prose-only) — parallel to v0.3 IDEATE-29 signed-attestation |
 | IDEATE-40 | P5 cross-link audit verifies edge↔messaging intra-v0.4 links both directions (refines ATELIER-114) | 05-02-01 | Cross-link audit checks edge→messaging AND messaging→edge links resolve; language-derived docs link to both new + existing domains where relevant |
 | IDEATE-41 | streams.md enumerates consumer-group rebalance strategies (refines ATELIER-100) | 02-02-03 | streams.md enumerates sticky/cooperative/eager — each with partition-stop-the-world cost + use-case fit (parallel to v0.3 IDEATE-30 drift-type enumeration) |
 | IDEATE-42 | Languages Coverage sub-table = exactly 16 rows (refines ATELIER-107) | 04-01-02, 06-01-01 | domain-coverage.md gains `## Languages Coverage` sub-table with exactly 16 rows; P6 review verifies the 16 rows match the derived docs' actual section-header traces |
 | IDEATE-43 | MANIFEST Cross-Cutting summary = "19 domains, 190 P-rules post-v0.4" (refines ATELIER-111) | 04-03-01 | MANIFEST Cross-Cutting summary line updates to exactly "19 domains, 190 P-rules post-v0.4" (IDEATE-18 carried forward) |
 ## v0.4 Decisions Logged (planning stage)
 | ID | Decision | Rationale | Confidence |
 |----|----------|-----------|------------|
 | D-070 | P3 Wave 2 splits into 2a/2b/2c/2d labels (by language) but all are the same dependency tier | 16 derived docs are all independent post-Wave-1; the 2a/2b/2c/2d labels organize by language, the executor schedules ≤5 concurrent per config.json. Avoids inventing a false dependency between languages (parallel to v0.3 D-049 i18n/compliance 2a/2b split) | 0.88 |
 | D-071 | P4 splits into 3 waves: matrix+coverage+ARCHITECTURE (Wave 1 sequential), review docs (Wave 2 parallel), MANIFEST (Wave 3 sequential after review docs settled) | Matrix is the arbiter (must be authoritative first per v0.2/v0.3 precedent); ARCHITECTURE.md follows coverage so the component tree reflects the matrix; review docs are independent of each other; MANIFEST lists file paths and totals, sequenced after anti-patterns content is settled (D-050 precedent — anti-patterns + manifest combine under lead-developer). Net 3 waves, respecting that ARCHITECTURE.md (ATELIER-117) is a new v0.4 deliverable that must follow the matrix/coverage extension | 0.83 |
 | D-072 | P6 splits into 2 waves: review (Wave 1 — edge-engineer + languages-engineer review their content per D-052 before removal), ship (Wave 2 — tag + merge + remove personas) | Phase-specific personas review their authored content one final time in Wave 1 (D-052 precedent — ensures D-061 edge boundary and D-063 apply-don't-derive are checked by the specialist), then are removed in Wave 2 (roster returns to 3 per D-067). Separating review from ship ensures the specialist review completes before the personas are deactivated | 0.84 |
 ## Assumptions Logged
 | # | Assumption | Confidence |
 |---|-----------|------------|
 | A-006 | The pre-existing matrix has exactly 170 P-rows (verified by grill G-014 direct count); appending 20 yields exactly 190 with no reconciliation needed. The IDEATE-34 "171-vs-170 discrepancy" was a false premise corrected by the grill (parallel to v0.3's matrix-extension assumption A-002) | 0.88 |
 | A-007 | The 4 new domain P-rules (edge + messaging) map to existing core C1–C8 without needing new core principles (core is stable at 8 — carried forward from A-002) | 0.95 |
 | A-008 | Wave 2 of P3 (16 derived docs) can be scheduled by the executor as 5-then-5-then-5-then-1 without a hard sub-wave barrier (carried forward from v0.3 A-003) | 0.90 |
 | A-009 | The Languages Coverage sub-table's 16 rows can be authored in P4 from the P3 derived docs' actual section-header traces (the RESEARCH-v0.4 sketch is the expected trace set; finalized when the derived docs are authored) | 0.85 |
 | A-010 | The edge↔messaging intra-milestone cross-links (edge/iot.md → messaging/queues.md) dangle at the P1 tag (v0.3.1) and resolve once P2 ships (v0.3.2); this is acceptable per D-053 vertical-slice integrity and verified bidirectionally in P5 (IDEATE-40) | 0.86 |
 ## Milestone Type: NFR
 Per branch-strategy.md, this is an NFR milestone (all phases are docs/chore — no `feat` code). NFR milestones produce progressive patches; the final patch (v0.3.6) IS the v0.4 deliverable. No separate minor tag.
@@ -164,6 +164,11 @@ NFR milestone: no separate minor tag. The final patch (v0.2.6) IS the v0.3 deliv
 | D-046 | i18n testing-i18n.md must map pseudo-locale testing to testing pyramid tiers | Avoids generic "test i18n" guidance; maps to unit (missing-key), integration (snapshot per locale), e2e (RTL coverage) | 0.78 |
 | D-047 | compliance evidence.md must include a fenced signed-attestation example (Cosign or in-toto) | Prose-only evidence guidance is weak; a fenced example demonstrates the principle concretely (P6 Evidence Collected Continuously) | 0.80 |
 | D-048 | ai-ml monitoring-drift.md must enumerate 3 drift types (data/concept/prediction) with a detection signal per type | Avoids conflating drift types; each has distinct detection signals and retraining triggers | 0.82 |
 | D-049 | P3 splits Wave 2 into 2a (i18n) + 2b (compliance) labels but both are the same dependency tier | 8 derived docs are all independent post-Wave-1; 2a/2b labels organize by domain, executor schedules ≤5 concurrent per config.json. Avoids inventing a false dependency between i18n and compliance | 0.88 |
 | D-050 | P4 Wave 2 runs exactly 5 concurrent tasks (at the max_concurrent_agents cap) | matrix, coverage, agent-checklist, peer-review-checklist, anti-patterns, manifest = 6 extensions, but anti-patterns (04-02-04) and manifest (04-02-05) are combined under lead-developer for manifest to sequence after anti-patterns content is settled. Net 5 concurrent slots | 0.82 |
 | D-051 | P5 ai-ml-reproducibility.md example authored by tech-writer with ml-engineer consultation (not ml-engineer primary) | ml-engineer is removed after P2 per PERSONAS.md; P5 examples are tech-writer territory. ml-engineer constraints are baked into the task must-have (P1/P2/P3/P5 demonstrated) so the constraint survives the persona | 0.80 |
 | D-052 | P6 review uses platform-engineer + ml-engineer for content review before removal | Phase-specific personas review their authored content one final time in P6 Wave 1, then are removed in 06-01-04. Ensures D-023 (ai-ml scope) and GitOps correctness are checked by the specialist before the roster returns to 3 | 0.84 |
 | D-053 | Vertical-slice integrity: each phase is independently shippable | P1 ships gitops-operators domain docs (matrix rows land in P4 — acceptable because the domain is self-consistent; matrix extension is the traceability closure, not a blocker for the domain's internal consistency). P3 ships 2 domains together (D-022). P4 closes traceability + manifest. P5 closes examples + cross-links. P6 ships the release | 0.86 |
 ## Cross-Project References
@@ -172,4 +177,51 @@ None yet. Atelier is a standalone docs framework.
 ## Milestone History
 - **v0.1** — Initial Framework (COMPLETE). 8 core principles, 11 domains, 110 domain principles, full matrix, 4+3 examples, 4 languages. Tag v0.0.7.
- **v0.2** — Infrastructure as Code + Kubernetes (COMPLETE). Adds 2 domains (20 new P-rules), matrix/review/examples integration. Tags v0.1.0–v0.1.5; v0.1.5 is the v0.2 release.
+- **v0.2** — Infrastructure as Code + Kubernetes (COMPLETE). Adds 2 domains (20 new P-rules), matrix/review/examples integration. Tags v0.1.0–v0.1.5; v0.1.5 is the v0.2 release.
 - **v0.3** — GitOps + Operators + AI/ML + i18n + Compliance (COMPLETE). Adds 4 domains (40 new P-rules), matrix 130→170 across 13→17 domains, review/checklist/anti-patterns integration, 4 new examples, MANIFEST examples/ listing (closes ESC-002). Tags v0.2.0–v0.2.6; v0.2.6 is the v0.3 release.
 - **v0.4** — Edge + Messaging + Language-Derived Docs (ACTIVE). Adds 2 new top-level domains (edge, messaging — 20 new P-rules), and extends `languages/` with derived docs per language (typescript, python, go, rust). Matrix 170→190 across 17→19 domains; review/checklist/anti-patterns extended; 4 new examples; MANIFEST gains a `languages/` section. Tags v0.3.0–v0.3.N; final patch IS the v0.4 deliverable.
 ## v0.4 — Edge + Messaging + Language-Derived Docs
 **Milestone type:** NFR (all phases produce docs — no `feat` runtime code)
 **Tag line:** v0.3.x (previous minor from v0.4)
 **Scope:** Extend the domain tree with two new top-level domains covering edge computing and messaging systems, plus extend `languages/` from first-principles-only stubs to full derived-doc trees. All content is docs-only markdown with illustrative code fences; no runtime/deployable artifacts. Closes the "Future Milestones" candidates in the v0.3 ROADMAP (edge domain + language-specific derived docs).
 ### New Domains
 - `domains/edge/` — edge computing domain (latency-bound, resource-constrained, geo-distributed)
  - `first-principles.md` — 10 edge principles (P1–P10)
  - Derived: `cdn.md`, `offline-first.md`, `iot.md`, `sync.md`
 - `domains/messaging/` — messaging systems domain (async, queues, streams, pub/sub)
  - `first-principles.md` — 10 messaging principles (P1–P10)
  - Derived: `queues.md`, `pubsub.md`, `streams.md`, `delivery-semantics.md`
 ### Language-Derived Extension
 Extend each `languages/*.md` from a single first-principles stub to a full derived-doc tree mirroring the domain structure (parallel to how `domains/` have first-principles + derived). Each language gets derived docs for the most-relevant domains (4 derived docs per language = 16 total), traced to that language's P-rules and to core C-rules.
 - `languages/typescript.md` (existing first-principles) + derived: `ts-types.md`, `ts-tooling.md`, `ts-async.md`, `ts-testing.md`
 - `languages/python.md` + derived: `py-types.md`, `py-tooling.md`, `py-async.md`, `py-testing.md`
 - `languages/go.md` + derived: `go-types.md`, `go-tooling.md`, `go-concurrency.md`, `go-testing.md`
 - `languages/rust.md` + derived: `rs-ownership.md`, `rs-tooling.md`, `rs-async.md`, `rs-testing.md`
 ### Cross-Domain Integration
 - Extend `matrix/principles-matrix.md` with 20 new P-rules → core C-rule mappings (10 per new domain; the language-derived docs trace to existing domain P-rules, not new matrix rows)
 - Extend `matrix/domain-coverage.md` with edge + messaging (row schema unchanged); add a `languages/` coverage sub-table listing each language's derived docs and the domains they trace to
 - Extend `review/agent-checklist.md`, `review/peer-review-checklist.md`, and `review/anti-patterns.md` with edge + messaging sections + a language-derived-docs section
 - Update `MANIFEST.md` with (a) the 2 new domains, (b) a new `languages/` section listing each language's first-principles + derived docs, (c) post-v0.4 totals (19 domains, 190 P-rules)
 - New examples (good + bad): `examples/good/edge-offline-sync.md`, `examples/good/messaging-idempotent-consumer.md`, `examples/bad/edge-sync-loop.md`, `examples/bad/messaging-shared-subscription.md`
 - Cross-links from new domains to existing `devops/`, `observability/`, `data/`, `concurrency/`, `kubernetes/`, `infrastructure-as-code/` domains; cross-links from language-derived docs to their traced domain docs
 ### Phase Plan (proposed, finalized in PLAN)
 - P0 Pre-Execution: spec, clarify, research, ideate, plan, grill
 - P1 Edge domain
 - P2 Messaging domain
 - P3 Language-Derived Extension (16 derived docs across 4 languages)
 - P4 Matrix + Review Integration (20 new mappings, manifest languages/ section, checklist parity)
 - P5 Examples + Cross-Links
 - P6 Final Review + Ship (IS the v0.4 release → tag v0.3.6)
 NFR milestone: no separate minor tag. The final patch (v0.3.6) IS the v0.4 deliverable.
@@ -133,38 +133,38 @@ All 35 requirements covered. 8 core principles, 11 domains, 110 domain principle
 | REQ-ID | Requirement | Priority | Phase | Status |
 |--------|-------------|----------|-------|--------|
-| ATELIER-60 | `domains/gitops-operators/first-principles.md` — 10 GitOps/operator principles (P1–P10) | P0 | 1 | pending |
+| ATELIER-60 | `domains/gitops-operators/first-principles.md` — 10 GitOps/operator principles (P1–P10) | P0 | 1 | covered |
-| ATELIER-61 | `domains/gitops-operators/argocd.md` — ArgoCD derived doc | P1 | 1 | pending |
+| ATELIER-61 | `domains/gitops-operators/argocd.md` — ArgoCD derived doc | P1 | 1 | covered |
-| ATELIER-62 | `domains/gitops-operators/flux.md` — Flux derived doc | P1 | 1 | pending |
+| ATELIER-62 | `domains/gitops-operators/flux.md` — Flux derived doc | P1 | 1 | covered |
-| ATELIER-63 | `domains/gitops-operators/operators.md` — Kubernetes Operators derived doc | P1 | 1 | pending |
+| ATELIER-63 | `domains/gitops-operators/operators.md` — Kubernetes Operators derived doc | P1 | 1 | covered |
-| ATELIER-64 | `domains/gitops-operators/progressive-delivery.md` — progressive delivery derived doc | P1 | 1 | pending |
+| ATELIER-64 | `domains/gitops-operators/progressive-delivery.md` — progressive delivery derived doc | P1 | 1 | covered |
-| ATELIER-65 | `domains/ai-ml/first-principles.md` — 10 AI/ML principles (P1–P10) | P0 | 2 | pending |
+| ATELIER-65 | `domains/ai-ml/first-principles.md` — 10 AI/ML principles (P1–P10) | P0 | 2 | covered |
-| ATELIER-66 | `domains/ai-ml/data-versioning.md` — data/model versioning derived doc | P1 | 2 | pending |
+| ATELIER-66 | `domains/ai-ml/data-versioning.md` — data/model versioning derived doc | P1 | 2 | covered |
-| ATELIER-67 | `domains/ai-ml/model-evaluation.md` — evaluation derived doc | P1 | 2 | pending |
+| ATELIER-67 | `domains/ai-ml/model-evaluation.md` — evaluation derived doc | P1 | 2 | covered |
-| ATELIER-68 | `domains/ai-ml/serving.md` — model serving derived doc | P1 | 2 | pending |
+| ATELIER-68 | `domains/ai-ml/serving.md` — model serving derived doc | P1 | 2 | covered |
-| ATELIER-69 | `domains/ai-ml/monitoring-drift.md` — monitoring/drift derived doc | P1 | 2 | pending |
+| ATELIER-69 | `domains/ai-ml/monitoring-drift.md` — monitoring/drift derived doc | P1 | 2 | covered |
-| ATELIER-70 | `domains/i18n/first-principles.md` — 10 i18n principles (P1–P10) | P0 | 3 | pending |
+| ATELIER-70 | `domains/i18n/first-principles.md` — 10 i18n principles (P1–P10) | P0 | 3 | covered |
-| ATELIER-71 | `domains/i18n/locale-resources.md` — locale resource management derived doc | P1 | 3 | pending |
+| ATELIER-71 | `domains/i18n/locale-resources.md` — locale resource management derived doc | P1 | 3 | covered |
-| ATELIER-72 | `domains/i18n/formatting.md` — formatting (dates/numbers/units) derived doc | P1 | 3 | pending |
+| ATELIER-72 | `domains/i18n/formatting.md` — formatting (dates/numbers/units) derived doc | P1 | 3 | covered |
-| ATELIER-73 | `domains/i18n/rtl-bidi.md` — RTL/bidi layout derived doc | P1 | 3 | pending |
+| ATELIER-73 | `domains/i18n/rtl-bidi.md` — RTL/bidi layout derived doc | P1 | 3 | covered |
-| ATELIER-74 | `domains/i18n/testing-i18n.md` — i18n testing derived doc | P1 | 3 | pending |
+| ATELIER-74 | `domains/i18n/testing-i18n.md` — i18n testing derived doc | P1 | 3 | covered |
-| ATELIER-75 | `domains/compliance/first-principles.md` — 10 compliance principles (P1–P10) | P0 | 3 | pending |
+| ATELIER-75 | `domains/compliance/first-principles.md` — 10 compliance principles (P1–P10) | P0 | 3 | covered |
-| ATELIER-76 | `domains/compliance/audit-logs.md` — audit logging derived doc | P1 | 3 | pending |
+| ATELIER-76 | `domains/compliance/audit-logs.md` — audit logging derived doc | P1 | 3 | covered |
-| ATELIER-77 | `domains/compliance/data-retention.md` — data retention derived doc | P1 | 3 | pending |
+| ATELIER-77 | `domains/compliance/data-retention.md` — data retention derived doc | P1 | 3 | covered |
-| ATELIER-78 | `domains/compliance/policy-as-code.md` — policy-as-code derived doc | P1 | 3 | pending |
+| ATELIER-78 | `domains/compliance/policy-as-code.md` — policy-as-code derived doc | P1 | 3 | covered |
-| ATELIER-79 | `domains/compliance/evidence.md` — evidence collection derived doc | P1 | 3 | pending |
+| ATELIER-79 | `domains/compliance/evidence.md` — evidence collection derived doc | P1 | 3 | covered |
-| ATELIER-80 | Extend `matrix/principles-matrix.md` with 40 new P-rules → core C-rule mappings (10 per new domain; review check: row count per domain = 10, each row ≥1 C-rule) | P0 | 4 | pending |
+| ATELIER-80 | Extend `matrix/principles-matrix.md` with 40 new P-rules → core C-rule mappings (10 per new domain; review check: row count per domain = 10, each row ≥1 C-rule) | P0 | 4 | covered |
-| ATELIER-81 | Extend `matrix/domain-coverage.md` with gitops-operators, ai-ml, i18n, compliance (row schema: domain, P-count, derived-doc-count, manifest-listed, status) | P1 | 4 | pending |
+| ATELIER-81 | Extend `matrix/domain-coverage.md` with gitops-operators, ai-ml, i18n, compliance (row schema: domain, P-count, derived-doc-count, manifest-listed, status) | P1 | 4 | covered |
-| ATELIER-82 | Extend `review/agent-checklist.md` with 4 new domain trigger sections | P1 | 4 | pending |
+| ATELIER-82 | Extend `review/agent-checklist.md` with 4 new domain trigger sections | P1 | 4 | covered |
-| ATELIER-83 | Extend `review/peer-review-checklist.md` with 4 new domain sections (parity with agent-checklist) | P1 | 4 | pending |
+| ATELIER-83 | Extend `review/peer-review-checklist.md` with 4 new domain sections (parity with agent-checklist) | P1 | 4 | covered |
-| ATELIER-84 | Extend `review/anti-patterns.md` with 4 new domain violations incl. orphaned P-rule + deployable example artifact | P1 | 4 | pending |
+| ATELIER-84 | Extend `review/anti-patterns.md` with 4 new domain violations incl. orphaned P-rule + deployable example artifact | P1 | 4 | covered |
-| ATELIER-85 | Update `MANIFEST.md` to list all new v0.3 documents (manifest authoritative) | P0 | 4 | pending |
+| ATELIER-85 | Update `MANIFEST.md` to list all new v0.3 documents (manifest authoritative) | P0 | 4 | covered |
-| ATELIER-86 | `examples/good/gitops-pr.md` + `examples/good/ai-ml-reproducibility.md` — 2 good examples (markdown with fenced code only) | P2 | 5 | pending |
+| ATELIER-86 | `examples/good/gitops-pr.md` + `examples/good/ai-ml-reproducibility.md` — 2 good examples (markdown with fenced code only) | P2 | 5 | covered |
-| ATELIER-87 | `examples/bad/i18n-string-concat.md` + `examples/bad/compliance-audit-log.md` — 2 named bad examples (each cites the P-rule breached) | P2 | 5 | pending |
+| ATELIER-87 | `examples/bad/i18n-string-concat.md` + `examples/bad/compliance-audit-log.md` — 2 named bad examples (each cites the P-rule breached) | P2 | 5 | covered |
-| ATELIER-88 | Cross-links from new domains to existing devops/security/observability/data/kubernetes/infrastructure-as-code domains (review check: every new derived doc ≥1 outbound cross-link to a MANIFEST-listed doc) | P1 | 5 | pending |
+| ATELIER-88 | Cross-links from new domains to existing devops/security/observability/data/kubernetes/infrastructure-as-code domains (review check: every new derived doc ≥1 outbound cross-link to a MANIFEST-listed doc) | P1 | 5 | covered |
-| ATELIER-89 | Final review passes (all v0.3 phases reviewed, audit clean) | P0 | 6 | pending |
+| ATELIER-89 | Final review passes (all v0.3 phases reviewed, audit clean) | P0 | 6 | covered |
-| ATELIER-90 | Milestone v0.3 released (tag v0.2.6, merged to main) | P0 | 6 | pending |
+| ATELIER-90 | Milestone v0.3 released (tag v0.2.6, merged to main) | P0 | 6 | covered |
-| ATELIER-91 | Add `examples/` directory listing to `MANIFEST.md` (pre-existing drift from v0.2 audit escalation ESC-002 note: examples/ unlisted; manifest is authoritative) | P1 | 4 | pending |
+| ATELIER-91 | Add `examples/` directory listing to `MANIFEST.md` (pre-existing drift from v0.2 audit escalation ESC-002 note: examples/ unlisted; manifest is authoritative) | P1 | 4 | covered |
 ## v0.3 Traceability Matrix
@@ -230,4 +230,166 @@ All 35 requirements covered. 8 core principles, 11 domains, 110 domain principle
 | IDEATE-03 (domain-coverage row schema) | ATELIER-81 + IDEATE-19 extend with C-rule coverage table update | ✓ extended |
 | IDEATE-07 (named bad examples cite P-rule breached) | ATELIER-87 + IDEATE-26 refine (two-breach example) | ✓ extended |
 | IDEATE-10/11/12 (decision/comparison tables) | IDEATE-21/22/23 transfer the pattern to 3 v0.3 derived docs | ✓ transferred |
-| v0.2 audit ESC-002 note (examples/ not in MANIFEST) | IDEATE-17 → ATELIER-91 | ✓ addressed |
+| v0.2 audit ESC-002 note (examples/ not in MANIFEST) | IDEATE-17 → ATELIER-91 | ✓ addressed |
 ## v0.4 Requirements — Edge + Messaging + Language-Derived Docs
 **Milestone type:** NFR (all phases produce docs)
 **Tag line:** v0.3.x (previous minor from v0.4)
 | REQ-ID | Requirement | Priority | Phase | Status |
 |--------|-------------|----------|-------|--------|
 | ATELIER-92 | `domains/edge/first-principles.md` — 10 edge principles (P1–P10) | P0 | 1 | pending |
 | ATELIER-93 | `domains/edge/cdn.md` — CDN/edge-cache derived doc (with edge-cache vs origin decision matrix) | P1 | 1 | pending |
 | ATELIER-94 | `domains/edge/offline-first.md` — offline-first derived doc | P1 | 1 | pending |
 | ATELIER-95 | `domains/edge/iot.md` — constrained-device/IoT derived doc | P1 | 1 | pending |
 | ATELIER-96 | `domains/edge/sync.md` — conflict/sync derived doc (CRDT/last-write-win decision matrix) | P1 | 1 | pending |
 | ATELIER-97 | `domains/messaging/first-principles.md` — 10 messaging principles (P1–P10) | P0 | 2 | pending |
 | ATELIER-98 | `domains/messaging/queues.md` — queue derived doc (with at-least-once/at-most-once/exactly-once comparison) | P1 | 2 | pending |
 | ATELIER-99 | `domains/messaging/pubsub.md` — pub/sub derived doc | P1 | 2 | pending |
 | ATELIER-100 | `domains/messaging/streams.md` — stream processing derived doc | P1 | 2 | pending |
 | ATELIER-101 | `domains/messaging/delivery-semantics.md` — delivery semantics derived doc (idempotency, ordering, dead-letter) | P1 | 2 | pending |
 | ATELIER-102 | `languages/typescript.md` extended + `languages/ts-types.md`, `ts-tooling.md`, `ts-async.md`, `ts-testing.md` — 4 TS derived docs | P1 | 3 | pending |
 | ATELIER-103 | `languages/python.md` extended + `languages/py-types.md`, `py-tooling.md`, `py-async.md`, `py-testing.md` — 4 Python derived docs | P1 | 3 | pending |
 | ATELIER-104 | `languages/go.md` extended + `languages/go-types.md`, `go-tooling.md`, `go-concurrency.md`, `go-testing.md` — 4 Go derived docs | P1 | 3 | pending |
 | ATELIER-105 | `languages/rust.md` extended + `languages/rs-ownership.md`, `rs-tooling.md`, `rs-async.md`, `rs-testing.md` — 4 Rust derived docs | P1 | 3 | pending |
 | ATELIER-106 | Extend `matrix/principles-matrix.md` with 20 new P-rules → core C-rule mappings (10 per new domain; review check: row count per domain = 10, each row ≥1 C-rule) — post-v0.4: 19 domains, 190 P-rules | P0 | 4 | pending |
 | ATELIER-107 | Extend `matrix/domain-coverage.md` with edge + messaging (per-domain row schema) + add `languages/` coverage sub-table (language → derived docs → traced domains) | P1 | 4 | pending |
 | ATELIER-108 | Extend `review/agent-checklist.md` with edge + messaging trigger sections + a language-derived-docs section | P1 | 4 | pending |
 | ATELIER-109 | Extend `review/peer-review-checklist.md` with edge + messaging + language-derived sections (parity with agent-checklist) | P1 | 4 | pending |
 | ATELIER-110 | Extend `review/anti-patterns.md` with edge + messaging violations incl. orphaned P-rule + deployable example artifact; pre-specify domain chaos anti-patterns (edge: sync conflict loop, TTL-less cache; messaging: shared subscription, blocking consumer) | P1 | 4 | pending |
 | ATELIER-111 | Update `MANIFEST.md` to list new v0.4 domains + new `languages/` section (first-principles + derived docs per language) + post-v0.4 totals | P0 | 4 | pending |
 | ATELIER-112 | `examples/good/edge-offline-sync.md` + `examples/good/messaging-idempotent-consumer.md` — 2 good examples (markdown with fenced code only) | P2 | 5 | pending |
 | ATELIER-113 | `examples/bad/edge-sync-loop.md` + `examples/bad/messaging-shared-subscription.md` — 2 named bad examples (each cites the P-rule breached) | P2 | 5 | pending |
 | ATELIER-114 | Cross-links from new domains to existing devops/observability/data/concurrency/kubernetes/infrastructure-as-code domains; cross-links from language-derived docs to their traced domain docs (review check: every new derived doc ≥1 outbound cross-link to a MANIFEST-listed doc) | P1 | 5 | pending |
 | ATELIER-115 | Final review passes (all v0.4 phases reviewed, audit clean) | P0 | 6 | pending |
 | ATELIER-116 | Milestone v0.4 released (tag v0.3.6, merged to main) | P0 | 6 | pending |
 | ATELIER-117 | Update `ARCHITECTURE.md` for v0.4 — extend the component tree with `edge/` + `messaging/` domain entries, add a `## v0.4 Domain Additions` section documenting the two new domains (boundary per D-061/D-062, derived docs, cross-link targets), and note the `languages/` derived-doc extension (D-063 apply-don't-derive). Closes architecture-tree drift: the file currently stops at v0.3 with no edge/messaging/languages-derived coverage (IDEATE-32). | P1 | 4 | pending |
 ## v0.4 Traceability Matrix
 | Phase | Requirements |
 |-------|-------------|
 | 0 (Pre-Execution) | (governance: spec, clarify, research, ideate, plan) |
 | 1 (Edge Domain) | ATELIER-92..ATELIER-96 |
 | 2 (Messaging Domain) | ATELIER-97..ATELIER-101 |
 | 3 (Language-Derived Extension) | ATELIER-102..ATELIER-105 |
 | 4 (Matrix + Review Integration) | ATELIER-106..ATELIER-111, ATELIER-117 |
 | 5 (Examples + Cross-Links) | ATELIER-112..ATELIER-114 |
 | 6 (Final Review + Ship) | ATELIER-115, ATELIER-116 |
 ## v0.4 Ideation Log
 **Generated:** 13 ideas (mechanical: 8, backend-enriched: 5)
 **Accepted:** 13 (all v0.4-scope, confidence ≥ 0.79, above 0.6 autonomy threshold → auto-accepted)
 **Skipped:** 0
 **New requirements added:** 1 (ATELIER-117 — ARCHITECTURE.md v0.4 update, from IDEATE-32)
 **Refinements to existing reqs:** 12 (IDEATE-31, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43)
 **Deferred to v0.5:** 0
 **Rejected:** 0
 | IDEATE-ID | Source | Category | Confidence | Decision | Mapped REQ |
 |-----------|--------|----------|------------|----------|------------|
 | IDEATE-31 | mechanical (MANIFEST drift, parallel to ESC-002) | drift | 0.86 | accepted → refines | ATELIER-111 (languages/ section must list 4 v0.1 stubs + 16 derived docs) |
 | IDEATE-32 | mechanical (ARCHITECTURE.md vs directory tree) | drift | 0.88 | accepted → new req | ATELIER-117 (update ARCHITECTURE.md for v0.4: component tree + v0.4 section) |
 | IDEATE-33 | mechanical (IDEATE-19 pattern transfer) | coverage | 0.90 | accepted → refines | ATELIER-107 (Core Principle Coverage C1–C8 → domains table update for edge + messaging) |
 | IDEATE-34 | mechanical (matrix row-count invariant) | drift | 0.82 | accepted → refines | ATELIER-106 (verify post-v0.4 = exactly 190; verify current = exactly 170 per grill G-014 — the "171 vs 170" claim was a false premise, corrected) |
 | IDEATE-35 | mechanical (D-069 vs req-row text) | spec | 0.89 | accepted → refines | ATELIER-100, ATELIER-101 (req text must reference streams + dead-letter comparison tables) |
 | IDEATE-36 | mechanical (RESEARCH anti-patterns vs ATELIER-110 text) | quality | 0.85 | accepted → refines | ATELIER-110 (language-derived-doc anti-patterns + .ts/.py/.go/.rs artifact-type extension) |
 | IDEATE-37 | backend-enriched (chaos anti-pattern enumeration) | chaos | 0.86 | accepted → refines | ATELIER-110 (include all 6 RESEARCH chaos anti-patterns: edge sync-loop/TTL-less/blocking-call; messaging shared-sub/blocking-consumer/unbounded-retry) |
 | IDEATE-38 | backend-enriched (IDEATE-28 pattern transfer) | improvement | 0.80 | accepted → refines | ATELIER-94 (offline-first.md maps offline write-queue patterns to testing pyramid tiers) |
 | IDEATE-39 | backend-enriched (IDEATE-29 pattern transfer) | improvement | 0.79 | accepted → refines | ATELIER-101 (delivery-semantics.md fenced idempotency-key dedup-store example, not prose-only) |
 | IDEATE-40 | backend-enriched (intra-milestone cross-link dependency) | chaos | 0.84 | accepted → refines | ATELIER-114 (P5 cross-link audit must verify edge↔messaging intra-v0.4 links resolve in both directions) |
 | IDEATE-41 | backend-enriched (IDEATE-30 pattern transfer) | improvement | 0.81 | accepted → refines | ATELIER-100 (streams.md enumerates consumer-group rebalance strategies: sticky/cooperative/eager) |
 | IDEATE-42 | mechanical (D-066 sub-table verifiability) | coverage | 0.87 | accepted → refines | ATELIER-107 (Languages Coverage sub-table = exactly 16 rows; P4 review check) |
 | IDEATE-43 | mechanical (IDEATE-18 pattern transfer) | coverage | 0.83 | accepted → refines | ATELIER-111 (MANIFEST Cross-Cutting summary = "19 domains, 190 P-rules post-v0.4") |
 ### Refinements Notes (applied to existing reqs at execute time, not changing req rows)
 - **ATELIER-106** (IDEATE-34, corrected by grill G-014): matrix coverage summary must read
  "post-v0.4: 19 domains, 190 P-rules" in BOTH the summary block and
  the per-domain section count (IDEATE-18 carried forward). P4 must
  verify the matrix contains exactly 190 P-rows (10 per domain × 19).
  The current matrix has **exactly 170 P-rows** (17 domains × 10,
  verified by grill G-014 direct count) — the IDEATE-34 "171 vs 170"
  claim was a FALSE PREMISE. P4 must verify 170 exists, append 20 rows,
  and confirm post-v0.4 count = 190. Do NOT hunt for or delete a phantom
  171st row.
 - **ATELIER-107** (IDEATE-33, IDEATE-42): the domain-coverage.md
  extension must update THREE artifacts: (a) the per-domain row-schema
  table with edge + messaging rows; (b) the "Core Principle Coverage"
  table (C1–C8 → domains) for the 2 new domains (C4 Locality grows by
  edge P1/P8 + messaging P7; C5 Reversibility grows by edge P2/P4/P7/P10
  + messaging P5/P8/P9; C7 Observability grows by edge P10 + messaging
  P8/P10; C8 Economy grows by edge P3/P9 + messaging P6; C2 grows by
  messaging P1/P2/P4) — parallel to v0.3 IDEATE-19; (c) the new
  `## Languages Coverage` sub-table with EXACTLY 16 rows (one per
  language-derived doc), verifiable as a P4 review check (no orphans,
  every derived doc traced to ≥1 domain P-rule).
 - **ATELIER-110** (IDEATE-36, IDEATE-37): the anti-patterns extension
  must include (a) the v0.4 deployable-artifact-types table extended to
  `.ts`/`.py`/`.go`/`.rs` standalone files (language-derived docs risk,
  per D-020 docs-only contract); (b) language-derived-doc anti-
  patterns: introducing a new P-rule (e.g., "TS P1"), orphan section
  with no domain P-rule trace, runtime standalone artifact, and
  first-principles stub gaining P1–P10 numbering; (c) all 6 RESEARCH-
  specified chaos anti-patterns, not just the 4 named in the req-row
  text — edge: sync conflict loop (P4), TTL-less cache (P6), blocking
  call on constrained device (P3/P5); messaging: shared subscription
  (P2/P3), blocking consumer (P6), unbounded retry / no DLQ (P5).
 - **ATELIER-100** (IDEATE-35, IDEATE-41): streams.md must include (a)
  the stream-platform comparison table (Kafka vs Kinesis vs
  Pulsar/NATS JetStream — ordering, partitioning, replay/retention,
  use-case fit) per D-069, AND (b) an enumeration of consumer-group
  rebalance strategies (sticky/cooperative vs eager) as a sub-
  comparison — each with partition-stop-the-world cost and use-case
  fit. Parallel to v0.3 IDEATE-30 (drift-type enumeration).
 - **ATELIER-101** (IDEATE-35, IDEATE-39): delivery-semantics.md must
  include (a) the dead-letter strategy comparison table (retry-count-
  limit vs TTL-with-backoff vs poison-queue vs DLQ+alert) per D-069,
  AND (b) a fenced idempotency-key dedup-store example (not prose-
  only) — parallel to v0.3 IDEATE-29 (fenced signed-attestation
  example). Prose-only delivery-semantics guidance is weak; a fenced
  consumer-with-dedup-store demonstrates P3 (idempotent consumers)
  concretely.
 - **ATELIER-111** (IDEATE-31, IDEATE-43): the new `## Languages`
  MANIFEST section must list the 4 v0.1 first-principles stubs
  (typescript.md, python.md, go.md, rust.md — closing the pre-existing
  drift since v0.1, parallel to ESC-002/ATELIER-91) AND the 16 new
  derived docs. The Cross-Cutting summary line must update to exactly
  "19 domains, 190 P-rules post-v0.4" (IDEATE-18 carried forward).
 - **ATELIER-114** (IDEATE-40): the P5 cross-link audit must verify
  edge↔messaging intra-v0.4 links resolve in BOTH directions, not only
  new→existing. Edge/iot.md (P1) cross-links messaging/queues.md (MQTT
  QoS parallels delivery semantics) and messaging/iot.md-style
  references — but messaging/ docs are authored in P2, so at the P1
  tag (v0.3.1) those links dangle. This is acceptable per vertical-
  slice integrity (D-053 — matrix/cross-links close in P4/P5), but the
  P5 audit (ATELIER-114) must explicitly check that every edge→
  messaging and messaging→edge link resolves by the P6 ship. The
  audit must also confirm language-derived docs (P3) link to BOTH
  new (edge/messaging) and existing domains where relevant (e.g.,
  ts-async.md → messaging/delivery-semantics; go-concurrency.md →
  concurrency/patterns + messaging/queues).
 - **ATELIER-94** (IDEATE-38): offline-first.md must map offline write-
  queue / conflict-detection patterns to the testing pyramid tiers
  (unit: conflict detection on a merge function; integration:
  reconnect reconcile against a local store; e2e: partition
  simulation with a fake network) — parallel to v0.3 IDEATE-28 (i18n
  pseudo-locale → pyramid). Avoids generic "test offline" guidance.
 ### Within-Project Pattern Transfer (v0.1 → v0.2 → v0.3 → v0.4) — verified
 | Prior Lesson | v0.4 Application | Status |
 |--------------|------------------|--------|
 | IDEATE-10/11/12 + 21/22/23 (decision/comparison tables in multi-tool derived docs) | D-069 + IDEATE-35 transfer the pattern to streams.md (Kafka/Kinesis/Pulsar) + delivery-semantics.md (dead-letter strategies) | ✓ transferred |
 | IDEATE-17 (manifest drift closure — examples/ unlisted → ATELIER-91) | IDEATE-31 → ATELIER-111 (languages/ unlisted since v0.1 — closes pre-existing drift on the same ESC-002 precedent) | ✓ transferred |
 | IDEATE-18 (coverage summary totals must update — string + per-domain count) | IDEATE-43 → ATELIER-111 (Cross-Cutting = "19 domains, 190 P-rules post-v0.4") + IDEATE-34 → ATELIER-106 (summary block) | ✓ transferred |
 | IDEATE-19 (Core Principle Coverage C1–C8 → domains table update) | IDEATE-33 → ATELIER-107 (C-rule coverage table extended for edge + messaging with the documented C-count shifts) | ✓ transferred |
 | IDEATE-20 (anti-patterns specificity: domain violations + artifact types) | IDEATE-36 + IDEATE-37 → ATELIER-110 (language-derived anti-patterns + .ts/.py/.go/.rs artifact types + all 6 chaos anti-patterns) | ✓ extended |
 | IDEATE-07 (named bad examples cite the P-rule breached) | ATELIER-113 + D-068 already cover (single-breach, cite specific P-rule) | ✓ carried forward |
 | IDEATE-08 (cross-link verification: every new derived doc ≥1 outbound) | ATELIER-114 already covers; IDEATE-40 extends to edge↔messaging intra-milestone links | ✓ extended |
 | IDEATE-13/14 (chaos: orphan P-rule + deployable artifact) | IDEATE-36 → ATELIER-110 (language orphan-section + standalone .ts/.py/.go/.rs runtime artifact) | ✓ extended |
 | IDEATE-28 (pseudo-locale testing → testing pyramid tiers) | IDEATE-38 → ATELIER-94 (offline write-queue testing → pyramid tiers) | ✓ transferred |
 | IDEATE-29 (fenced signed-attestation example, not prose-only) | IDEATE-39 → ATELIER-101 (fenced idempotency-key dedup-store example) | ✓ transferred |
 | IDEATE-30 (drift-type enumeration: data/concept/prediction + signal each) | IDEATE-41 → ATELIER-100 (rebalance-strategy enumeration: sticky/cooperative/eager + cost each) | ✓ transferred |
@@ -86,11 +86,11 @@ NFR milestone: no separate minor tag. The final patch (v0.1.5) IS the v0.2 deliv
 |-------|------|------|--------|------------------|
 | 0 | Pre-Execution | docs | complete | Spec, clarify, research, ideate, plan, PERSONAS.md (extends platform-engineer, adds ml-engineer) — shipped v0.2.0 |
 | 1 | GitOps + Operators Domain | docs | complete | domains/gitops-operators/{first-principles, argocd, flux, operators, progressive-delivery}.md — shipped v0.2.1 |
-| 2 | AI/ML Domain | docs | pending | domains/ai-ml/{first-principles, data-versioning, model-evaluation, serving, monitoring-drift}.md |
+| 2 | AI/ML Domain | docs | complete | domains/ai-ml/{first-principles, data-versioning, model-evaluation, serving, monitoring-drift}.md — shipped v0.2.2 |
-| 3 | i18n + Compliance Domains | docs | pending | domains/i18n/{first-principles, locale-resources, formatting, rtl-bidi, testing-i18n}.md, domains/compliance/{first-principles, audit-logs, data-retention, policy-as-code, evidence}.md |
+| 3 | i18n + Compliance Domains | docs | complete | domains/i18n/{first-principles, locale-resources, formatting, rtl-bidi, testing-i18n}.md, domains/compliance/{first-principles, audit-logs, data-retention, policy-as-code, evidence}.md — shipped v0.2.3 |
-| 4 | Matrix + Review Integration | docs | pending | matrix/principles-matrix.md (+40 mappings), matrix/domain-coverage.md (incl. C-rule coverage table update), review/{agent-checklist, peer-review-checklist, anti-patterns}.md, MANIFEST.md (+ examples/ listing per ATELIER-91) |
+| 4 | Matrix + Review Integration | docs | complete | matrix/principles-matrix.md (+40 mappings), matrix/domain-coverage.md (incl. C-rule coverage table update), review/{agent-checklist, peer-review-checklist, anti-patterns}.md, MANIFEST.md (+ examples/ listing per ATELIER-91) — shipped v0.2.4 |
-| 5 | Examples + Cross-Links | docs | pending | examples/good + examples/bad for 4 domains, cross-links to devops/security/observability/data/k8s/iac |
+| 5 | Examples + Cross-Links | docs | complete | examples/good + examples/bad for 4 domains, cross-links to devops/security/observability/data/k8s/iac — shipped v0.2.5 |
-| 6 | Final Review + Ship | docs | pending | Review passed, audit clean, milestone merged to main, tag v0.2.6 |
+| 6 | Final Review + Ship | docs | complete | Review passed, audit clean, milestone merged to main, tag v0.2.6 |
 ## v0.3 Phase Tag Mapping
@@ -119,7 +119,44 @@ NFR milestone: no separate minor tag. The final patch (v0.2.6) IS the v0.3 deliv
 ## Future Milestones
- **v0.4** (candidates): `domains/edge/`, `domains/quantum/`, language-specific derived docs, tooling adapters (linters), translation/localization of framework docs.
+- **v0.4** (ACTIVE): `domains/edge/`, `domains/messaging/`, `languages/`-derived docs (typescript/python/go/rust → 4 derived docs each).
 - **v0.5** (candidates): `domains/quantum/`, tooling adapters (linters), translation/localization of framework docs.
 ## Milestone: v0.4 — Edge + Messaging + Language-Derived Docs (ACTIVE)
 **Milestone type:** NFR (all phases produce docs — no `feat` code)
 **Tag line:** v0.3.x (previous minor from v0.4)
 **Phases:** P0 (pre-execution) + P1–P5 (execution) + P6 (final review+ship)
 | Phase | Name | Type | Status | Key Deliverables |
 |-------|------|------|--------|------------------|
 | 0 | Pre-Execution | docs | complete | Spec, clarify, research, ideate, plan, PERSONAS.md (adds edge-engineer + languages-engineer phase-specific personas) — shipped v0.3.0 |
 | 1 | Edge Domain | docs | complete | domains/edge/{first-principles, cdn, offline-first, iot, sync}.md — shipped v0.3.1 |
 | 2 | Messaging Domain | docs | pending | domains/messaging/{first-principles, queues, pubsub, streams, delivery-semantics}.md |
 | 3 | Language-Derived Extension | docs | pending | languages/ × 4 → first-principles + 4 derived docs each (16 derived docs) |
 | 4 | Matrix + Review Integration | docs | pending | matrix/principles-matrix.md (+20 mappings, 170→190), matrix/domain-coverage.md (+ languages/ sub-table), review/{agent-checklist, peer-review-checklist, anti-patterns}.md, MANIFEST.md (languages/ section) |
 | 5 | Examples + Cross-Links | docs | pending | examples/good + examples/bad for 2 domains, cross-links to devops/observability/data/concurrency/kubernetes/infrastructure-as-code + language→domain links |
 | 6 | Final Review + Ship | docs | pending | Review passed, audit clean, milestone merged to main, tag v0.3.6 |
 ## v0.4 Phase Tag Mapping
 Per branch-strategy.md, milestone `v0.4` tags run on the `v0.3.x` patch line:
 | Phase | Tag | Notes |
 |-------|-----|-------|
 | P0 | v0.3.0 | Pre-execution release |
 | P1 | v0.3.1 | Edge domain |
 | P2 | v0.3.2 | Messaging domain |
 | P3 | v0.3.3 | Language-derived extension |
 | P4 | v0.3.4 | Matrix + review integration |
 | P5 | v0.3.5 | Examples + cross-links |
 | P6 | v0.3.6 | Final review + ship — **IS the v0.4 milestone release** |
 NFR milestone: no separate minor tag. The final patch (v0.3.6) IS the v0.4 deliverable.
 ## v0.4 Ideation Outcome
 (Populated during P0 IDEATE stage.)
 ## Success Criteria
@@ -3,7 +3,7 @@
    {
      "slug": "atelier",
      "name": "Atelier",
-      "milestone": "v0.3",
+      "milestone": "v0.4",
      "status": "active"
    }
  ],
@@ -36,13 +36,43 @@
 | DevOps        | ✓                | ci-cd, environments |
 | Infrastructure as Code | ✓      | terraform, opentofu, state, modules |
 | Kubernetes    | ✓                | workloads, networking, storage, rbac, helm, kustomize |
 | GitOps + Operators | ✓           | argocd, flux, operators, progressive-delivery |
 | AI / ML       | ✓                | data-versioning, model-evaluation, serving, monitoring-drift |
 | i18n          | ✓                | locale-resources, formatting, rtl-bidi, testing-i18n |
 | Compliance    | ✓                | audit-logs, data-retention, policy-as-code, evidence |
 ## Examples
 > Examples are illustrative markdown with fenced code only (no standalone runtime artifacts per D-020 / D-025). The `examples/` directory listing closes the v0.2 ESC-002 drift (IDEATE-17, ATELIER-91). P5 authored the v0.3 examples and promoted all entries from `pending` to `✓` (verified — every listed file exists).
 | Path | Status | Notes |
 |------|--------|-------|
 | `examples/good/` | ✓ | Good-example directory — 8 examples (v0.1 + v0.2 + v0.3) |
 | `examples/bad/` | ✓ | Bad-example directory — 7 examples (v0.1 + v0.2 + v0.3) |
 | `examples/good/api-endpoint.md` | ✓ | v0.1 example — good REST endpoint |
 | `examples/good/react-component.md` | ✓ | v0.1 example — good React component |
 | `examples/good/db-schema.md` | ✓ | v0.1 example — good DB schema |
 | `examples/good/error-handler.md` | ✓ | v0.1 example — good error handler |
 | `examples/bad/god-object.md` | ✓ | v0.1 example — bad god object |
 | `examples/bad/silent-error.md` | ✓ | v0.1 example — bad silent error |
 | `examples/bad/leaky-abstraction.md` | ✓ | v0.1 example — bad leaky abstraction |
 | `examples/good/terraform-module.md` | ✓ | v0.2 example — good IaC module |
 | `examples/good/k8s-deployment.md` | ✓ | v0.2 example — good k8s deployment |
 | `examples/bad/terraform-unlocked-state.md` | ✓ | v0.2 example — bad unlocked state |
 | `examples/bad/k8s-bare-pod-no-resources.md` | ✓ | v0.2 example — bad bare pod |
 | `examples/good/gitops-pr.md` | ✓ | v0.3 example — good GitOps PR |
 | `examples/good/ai-ml-reproducibility.md` | ✓ | v0.3 example — good reproducible training run |
 | `examples/bad/i18n-string-concat.md` | ✓ | v0.3 example — bad i18n string concat |
 | `examples/bad/compliance-audit-log.md` | ✓ | v0.3 example — bad audit log (P1 + P9 breaches) |
 > **Note:** The `examples/` section was established in P4 with entries pre-listed as `pending P5`. P5 authored the 4 v0.3 examples and promoted all entries to `✓` after verifying every listed file exists on disk. The manifest remains authoritative — unlisted = not part of framework.
 ## Cross-Cutting
 | Document                          | Purpose                          |
 |-----------------------------------|----------------------------------|
-| `matrix/principles-matrix.md`     | Maps domain → core principles (13 domains, 130 P-rules post-v0.2) |
+| `matrix/principles-matrix.md`     | Maps domain → core principles (17 domains, 170 P-rules post-v0.3) |
-| `matrix/domain-coverage.md`       | Maps core → domains; per-domain coverage |
+| `matrix/domain-coverage.md`       | Maps core → domains; per-domain coverage (incl. v0.3 Core Principle Coverage) |
-| `review/agent-checklist.md`       | Pre-completion agent checklist (incl. IaC + k8s triggers) |
+| `review/agent-checklist.md`       | Pre-completion agent checklist (incl. IaC + k8s + gitops + ai-ml + i18n + compliance triggers) |
-| `review/peer-review-checklist.md` | Human peer-review checklist (incl. IaC + k8s sections) |
+| `review/peer-review-checklist.md` | Human peer-review checklist (incl. IaC + k8s + gitops + ai-ml + i18n + compliance sections) |
-| `review/anti-patterns.md`         | Catalog of violations (incl. IaC + k8s + chaos anti-patterns) |
+| `review/anti-patterns.md`         | Catalog of violations (incl. IaC + k8s + gitops + ai-ml + i18n + compliance + v0.3 chaos anti-patterns) |
@@ -0,0 +1,165 @@
 # Audit Logs — Derived Rules
 > Derives from `domains/compliance/first-principles.md`. Covers P1
 > (Audit Logs are Append-Only), P2 (Every Significant Action is
 > Logged), P7 (Identity is Attributable), P9 (Secrets Redacted in
 > Audit), and P10 (Compliance Posture Observable). Referenced by
 > `data-retention.md` (retention applies to audit logs themselves)
 > and `evidence.md` (audit logs are evidence).
 ## Audit Logs are Append-Only (P1 Audit Logs are Append-Only)
 - An audit record is immutable once written. The storage substrate
  enforces this; policy alone does not. Write-once, append-only
  sinks (WORM buckets, immutable log streams, hash-chained ledgers)
  are the mechanism.
 - Deletion or mutation of an audit record is itself an auditable
  incident. The tampering is the signal, not just the underlying
  event. A system that allows `DELETE FROM audit_log` is a system
  whose audit log is a draft.
 - The append-only guarantee is testable: attempt to write, then
  attempt to overwrite, then attempt to delete. If the overwrite or
  delete succeeds, the guarantee is absent and the design is a
  violation.
 ## Structured Audit Events (P2 Every Significant Action is Logged)
 - The set of auditable actions is defined a priori, in code, before
  the action ships. The catalog is versioned and reviewed. An
  auditable action with no log line is a violation, not a gap to
  backfill later.
 - Audit events are structured (JSON / protobuf / a typed schema),
  not prose. A prose log line ("user logged in") is unqueryable and
  unaggregatable; a structured event is both. The event schema is
  the contract between the producer and the audit pipeline.
  ```
  {
    "timestamp": "2024-11-07T15:03:22Z",
    "event": "auth.login",
    "actor": { "kind": "user", "id": "u_8f3a", "session": "s_12b9" },
    "action": "succeeded",
    "target": { "kind": "service", "id": "billing-api" },
    "source": { "ip": "203.0.113.42", "region": "us-east-1" },
    "request_id": "req_91c2",
    "version": "audit-schema/v2"
  }
  ```
 - The catalog of significant actions typically includes:
  authentication (success and failure), authorization decisions
  (allow and deny), data access (read, write, delete), configuration
  changes, policy changes, retention executions, and admin
  operations. The exact set is declared per system; the discipline
  is that it is declared.
 ## Cloud Audit Log Conventions (Prior Art, Abstracted)
 - AWS CloudTrail, Google Cloud Audit Logs, and Azure Activity Log
  share a common shape: immutable, time-ordered, queryable, with
  actor / action / target / source / result fields. Atelier's
  audit-logs doc adopts the shape, not the vendor.
 - The shape is the contract; the sink is the implementation. A
  self-hosted audit log that follows the same shape composes with
  the same tooling (SIEM, query engines, evidence exporters) as the
  cloud vendors'.
 ## Queryability (P10 Compliance Posture Observable)
 - An audit log that cannot be queried is an audit log that cannot be
  used. Queryability is a first-class design goal: the event schema
  is typed, fields are indexed, and the common queries (who acted on
  what when, what failed, what was denied) are cheap.
 - "Who did X between T1 and T2" must be a single query, not a
  forensics project. If the query requires a custom script per
  investigation, the audit log is structured for storage, not for
  use — a C7 (Observability) violation.
 ## Identity is Attributable (P7 Identity is Attributable)
 - Every audit event records the authenticated principal that acted —
  not a shared account, not a generic service, not "admin." The
  actor field is populated at the time of the action from the
  authenticated session, not resolved after the fact.
 - A shared account in the actor field breaks accountability: an
  event attributed to `svc-deploy` could be any of ten engineers.
  This is the compliance angle on `domains/security/authorization.md`
  and `domains/kubernetes/rbac.md`: bind actions to unique
  principals, not to roles many can assume.
 - Machine-to-machine actions record the workload identity (a service
  account, a signed instance identity), not a human — but the
  identity is still unique and attributable to a deployable unit.
 ## Redaction at the Boundary (P9 Secrets Redacted in Audit)
 - Audit logs must not leak secrets, credentials, tokens, or PII.
  Redaction is structural: applied at the logging boundary, before
  the record is written to the append-only sink — not opportunistic
  scrubbing after the fact. Once a secret is in an append-only log,
  the remediation is expensive (rotate, rewrite access scope), so
  redaction-at-source is the only sound position.
 - The redaction policy is itself auditable: which fields are
  redacted, by what rule, in which event type. A redaction rule
  that lives in someone's head is a P9 violation waiting to happen.
  ```
  // before redaction (DO NOT LOG)
  {
    "event": "config.read",
    "target": { "kind": "secret", "id": "db-password" },
    "value": "p@ssw0rd-plaintext-leaked"   // VIOLATION
  }
  // after structural redaction
  {
    "event": "config.read",
    "target": { "kind": "secret", "id": "db-password" },
    "value": "[REDACTED:secret]",
    "redaction": "secret-value-policy/v1"
  }
  ```
 - Never log request bodies, response bodies, headers like
  `Authorization`, or environment variables that may carry secrets.
  Log the *fact* of the action, not the *content* of the secret.
 ## Retention of Audit Logs Themselves
 - Audit logs are subject to retention policy (cross `data-
  retention.md`), but the floor is set by the accountability need,
  not by storage economy. An audit log deleted before its retention
  period is a P1 violation dressed as a P3 action.
 - The retention rule for audit logs is itself logged (meta-audit):
  when an audit log segment ages out and is deleted, the deletion is
  recorded in a higher-tier audit log with the rule that authorized
  it. The chain is observable end to end.
 ## What Violates Audit-Log Discipline
 | Violation | Principle |
 |-----------|-----------|
 | Audit log on a mutable filesystem with no write-once protection | P1 Audit Logs are Append-Only |
 | `DELETE FROM audit_log WHERE timestamp < ...` as routine cleanup | P1 Audit Logs are Append-Only |
 | An auth-success event with no audit record | P2 Every Significant Action is Logged |
 | A prose log line ("user did a thing") instead of a structured event | P2 Every Significant Action is Logged |
 | A shared `admin` account as the actor in audit events | P7 Identity is Attributable |
 | An `Authorization: Bearer <token>` header logged in plaintext | P9 Secrets and Sensitive Data are Redacted in Audit |
 | A redaction rule applied inconsistently across event types | P9 Secrets and Sensitive Data are Redacted in Audit |
 | "Who did X?" requires a custom forensics script per investigation | P10 Compliance Posture is Observable |
 | An audit log segment deleted with no meta-audit record | P1 Audit Logs are Append-Only |
 ## Relationship to Other Domains
 - `domains/observability/logging.md` — audit logs are structured
  logging with an append-only guarantee; the logging primitives
  (levels, structured fields, correlation IDs) compose here.
 - `domains/security/authorization.md` — the actor in an audit event
  is the principal the authorization layer authenticated.
 - `domains/security/secrets.md` — redaction at the logging boundary
  is the audit-side complement of secret management.
 - `domains/compliance/data-retention.md` — retention policy applies
  to audit logs; the audit log's own deletion is meta-audited.
 - `domains/compliance/evidence.md` — audit logs are a primary
  evidence artifact; the append-only guarantee is what makes them
  admissible.
 - `domains/kubernetes/rbac.md` — workload identity in audit events
  derives from the RBAC principal that acted.
@@ -0,0 +1,159 @@
 # Data Retention — Derived Rules
 > Derives from `domains/compliance/first-principles.md`. Covers P3
 > (Retention is Policy, Not Storage) and the data-shape angle on P8
 > (Subject Access is Honored). Referenced by `audit-logs.md`
 > (retention applies to audit logs) and `evidence.md` (evidence has
 > a retention lifecycle). Framework-agnostic per D-024 — no
 > regulation-specific retention periods.
 ## Retention is Policy, Not Storage (P3 Retention is Policy, Not Storage)
 - Data lifetime is declared and enforced as policy, in code — not
  left to the storage layer's defaults. The policy names what data
  class is retained for how long, what action fires at end-of-life
  (delete, archive, anonymize), and what exception path exists (a
  legal hold suspends deletion).
 - Deletion at end-of-life is a feature, not a failure. A system that
  cannot delete on schedule is a system that over-retains, which is
  the symmetric violation of a system that under-retains. Both are
  P3 violations; the policy is the arbiter.
 - "We kept it because the bucket was cheap" is a violation. "We
  deleted it because the policy said to" is correct. Cost does not
  override policy; policy is the contract.
 ## Retention Policy as Code
 - Retention rules live as code: lifecycle rules on the storage
  layer, scheduled deletion jobs, tiered storage transitions, and
  anonymization transforms. The code is versioned, reviewed, and
  auditable. A retention rule in a spreadsheet is a wishlist; the
  same rule in a reviewed, deployable lifecycle policy is a control.
  ```
  // illustrative lifecycle policy (abstracted, no vendor DSL)
  // object-storage lifecycle
  {
    "rules": [
      {
        "name": "user-events-90d",
        "match": { "prefix": "events/" },
        "transitions": [
          { "after": "30d", "to": "tier-cold" },
          { "after": "90d", "action": "delete" }
        ]
      },
      {
        "name": "audit-log-7y",
        "match": { "prefix": "audit/" },
        "transitions": [
          { "after": "365d", "to": "tier-archive" },
          { "after": "2555d", "action": "delete" }
        ],
        "legal_hold": "suspends-action"
      }
    ]
  }
  ```
 - The retention policy is itself auditable: which rule fired when,
  against which objects, with what result. The deletion events are
  logged (`audit-logs.md`) — deletion is a significant action.
 ## Retention vs. Backup — The Distinction
 - A **backup** is a recovery mechanism: it exists to restore data
  after loss. A **retention rule** is a deletion mechanism: it
  exists to remove data at end-of-life. Conflating them produces
  data that survives both the deletion policy and the disaster —
  which is the opposite of compliance.
 - A backup is governed by a recovery-point / recovery-time objective;
  a retention rule is governed by a lifetime. They are independent
  contracts. A backup that is also the retention store is a store
  where nothing is ever deleted, which is a P3 violation.
 - A legal hold suspends retention deletion for a defined data set
  (e.g. data under investigation). The hold is itself a policy
  action, auditable and time-bounded, not a manual override.
 ## Retention is Distinct per Data Class
 - Different data classes have different lifetimes. The retention
  policy enumerates the classes and their rules; it does not apply
  one number to everything. Typical classes (the names are
  abstract; the periods are policy decisions, not regulation-
  specific):
  - **Audit logs** — long, often multi-year, governed by
    accountability needs (`audit-logs.md`).
  - **User-generated content** — tied to the user's account
    lifetime; deletion follows account deletion (cross P8 Subject
    Access).
  - **Telemetry / metrics** — short, governed by observability need
    (`domains/observability/metrics.md`); high-resolution data ages
    to downsampled aggregates.
  - **Evidence artifacts** — tied to the audit cycle
    (`evidence.md`); the cycle ends, the evidence ages out.
 - A single retention rule for "all data" is a C3 (Simplicity)
  violation of the wrong kind: it is simpler than the requirement
  allows.
 ## Subject Access is Honored (P8 Subject Access is Honored)
 - Data-subject rights — access (what do we have on this subject),
  export (in a portable form), deletion (and prove it), correction
  — are operations with defined contracts and audit trails, not
  ad-hoc tickets. The system implements them as first-class
  operations; a subject-access request that requires a forensics
  team is a correctness defect.
 - Retention and subject access interact at deletion: a subject
  deletion request fires the deletion policy for that subject's
  data, the deletion is audited, and the proof of deletion is
  returned to the subject (and recorded). A subject deletion that
  skips the audit is a P8 violation dressed as a P3 success.
 - Cross `domains/data/schema-design.md`: subject access is only
  computable if the schema tags which records belong to which
  subject. A schema with no subject linkage cannot honor a subject
  request — it cannot find the data to delete.
 ## Retention Migration Discipline
 - Retention rules change. When the policy changes (a class's
  lifetime shortens or lengthens), the change is a migration: the
  new rule applies to data ingested after the cutover, and a
  backfill applies the new rule to existing data where applicable.
  Cross `domains/data/migrations.md` for the schema-lifecycle
  discipline this mirrors.
 - A retention rule change that is not versioned, not reviewed, and
  not backfilled is a P3 violation: the policy is not actually the
  policy if the storage layer does not reflect it.
 ## What Violates Retention Discipline
 | Violation | Principle |
 |-----------|-----------|
 | Data kept indefinitely because "storage is cheap" | P3 Retention is Policy, Not Storage |
 | A retention rule in a spreadsheet, not in code | P3 Retention is Policy, Not Storage |
 | A backup bucket used as the retention store (nothing ever deletes) | P3 Retention is Policy, Not Storage |
 | A single retention period applied to all data classes | P3 Retention is Policy, Not Storage |
 | A subject deletion with no audit record of the deletion | P8 Subject Access is Honored |
 | A subject-access request that requires a forensics team | P8 Subject Access is Honored |
 | A schema with no subject linkage (cannot find data to delete) | P8 Subject Access is Honored |
 | A legal hold applied ad hoc, not as a policy action | P3 Retention is Policy, Not Storage |
 | A retention rule change with no backfill to existing data | P3 Retention is Policy, Not Storage |
 ## Relationship to Other Domains
 - `domains/data/schema-design.md` — retention requires the schema
  to tag data class and subject linkage; subject access is only
  computable over a schema that supports it.
 - `domains/data/migrations.md` — retention rule changes are
  migrations; the discipline (version, review, backfill) mirrors
  schema migrations.
 - `domains/compliance/audit-logs.md` — audit logs have their own
  retention floor; deletion of an audit segment is meta-audited.
 - `domains/compliance/evidence.md` — evidence artifacts have a
  retention lifecycle tied to the audit cycle.
 - `domains/observability/metrics.md` — telemetry retention is
  governed by observability need; high-res data ages to aggregates.
 - `domains/security/secrets.md` — secrets have a retention lifecycle
  tied to rotation; a secret past its rotation date is overdue, not
  retained.
@@ -0,0 +1,175 @@
 # Evidence — Derived Rules
 > Derives from `domains/compliance/first-principles.md`. Covers P6
 > (Evidence is Collected Continuously), P5 (Policy is a Gate, so
 > decisions are evidence), P7 (Identity Attributable, so evidence
 > has provenance), and P10 (Posture Observable, so evidence is
 > queryable). Referenced by `audit-logs.md` (logs are evidence)
 > and `data-retention.md` (evidence has a lifecycle).
 ## Evidence is Collected Continuously (P6 Evidence is Collected Continuously)
 - Evidence of compliance — logs, configs, scans, attestations,
  policy decisions, access reviews — is gathered as a byproduct of
  operation, not assembled manually at audit time. The audit-time
  scramble is the anti-pattern: it is expensive, it is incomplete,
  and it produces evidence that is reconstructed rather than
  recorded.
 - Continuous evidence collection means the audit packet is a query
  over already-collected artifacts, not a forensic reconstruction.
  The auditor asks "show me the access reviews for Q3" and the
  answer is a query against the evidence store, not a six-week
  project.
 - This is the compliance angle on `domains/observability/tracing.md`
  for distributed evidence (a trace spans the request that produced
  the evidence) and `domains/observability/metrics.md` for posture
  signals (a metric is a continuous evidence stream).
 ## Evidence is a Byproduct, Not a Deliverable
 - Evidence collected as a byproduct is trustworthy: it records what
  happened, when it happened, recorded by the system that did it.
  Evidence assembled at audit time is less trustworthy: it records
  what someone remembered to write down, when they wrote it, after
  the fact.
 - The mechanism: every significant action (`audit-logs.md`) emits
  its record to an evidence store; every policy decision
  (`policy-as-code.md`) emits its decision; every deployment emits
  its signed attestation; every access review emits its result. The
  store is append-only (`audit-logs.md` P1), queryable (P10), and
  retention-bound (`data-retention.md`).
 ## Provenance and Identity (P7 Identity is Attributable)
 - Evidence has provenance: which system produced it, when, from what
  input. An evidence artifact with no provenance is anecdote, not
  evidence — it cannot be attributed to a source, so it cannot be
  trusted.
 - Provenance includes the identity of the producer (a workload
  identity, a service account) and the chain of custody (who has
  had access to the artifact since it was produced). Cross
  `domains/security/authorization.md`: the producer's identity is
  authenticated, not assumed.
 ## Signed Attestations (IDEATE-29)
 - A signed attestation is evidence with a cryptographic signature
  binding the artifact to its producer. The signature is the
  provenance: it can be verified independently of the producer, and
  it cannot be forged without the producer's key. Cross
  `domains/security/supply-chain.md` for the supply-chain angle.
 - Cosign (Sigstore) and in-toto are the canonical patterns: a
  builder signs an artifact (container image, deployable, evidence
  bundle) at production time; a verifier checks the signature at
  consumption time. The signature is the evidence that the artifact
  came from where it claims to have come from.
 - **Illustrative signed attestation (Cosign / Sigstore format, NOT
  a real signature — illustrative only, no live keys):**
  ```
  // Cosign attest — bind an attestation to an image digest
  // (illustrative; not a real signature)
  $ cosign attest --type spdxjson \
      --predicate sbom.spdx.json \
      my-registry/app@sha256:5a3e1c...f9b2
  // The attestation is stored as a signature in the registry,
  // bound to the image digest. The payload is a DSSE envelope:
  {
    "payloadType": "application/vnd.in-toto+json",
    "payload": "eyJfdHlwZSI6ImF0dGVzdGF0aW9uIn0...",
    "signatures": [
      {
        "sig": "MEUCIQDx...illustrative-base64-signature...==",
        "keyid": "cosign-key-2024-q4"
      }
    ]
  }
  // The decoded payload (an in-toto statement binding the
  // attestation to the image digest):
  {
    "_type": "https://in-toto.io/Statement/v0.1",
    "predicateType": "https://spdx.dev/Document",
    "subject": [
      {
        "name": "my-registry/app",
        "digest": { "sha256": "5a3e1c...f9b2" }
      }
    ],
    "predicate": {
      "SPDXID": "SPDXRef-DOCUMENT",
      "creationInfo": {
        "created": "2024-11-07T15:03:22Z",
        "creators": ["Tool: atelier-build-pipeline"]
      }
    }
  }
  // Verification (independent of the producer):
  $ cosign verify-attestation --type spdxjson \
      --certificate-identity-regexp '.*atelier-build.*' \
      my-registry/app@sha256:5a3e1c...f9b2
  // Verification succeeded for: my-registry/app@sha256:5a3e1c...f9b2
  // SBOM attestation found for subject
  ```
 - The attestation is illustrative — the signatures and digests are
  not real. The shape (DSSE envelope, in-toto statement, subject +
  predicate, verify-by-identity) is what evidence-as-attestation
  looks like. A real attestation carries a real signature from a
  real key held by the builder.
 ## Audit-Ready Export
 - The evidence store is queryable at any time, not only at audit
  time. The audit packet is a query (a date range, a data class, a
  subject) over the store; the export is a dump of the matching
  artifacts with their provenance and signatures.
 - An audit-ready export that requires six weeks of forensics is a
  P6 violation dressed as a success: the evidence was not collected
  continuously, it was reconstructed. The export should be a query
  that runs in minutes, not a project that runs for weeks.
 ## Evidence Lifecycle
 - Evidence has a retention lifecycle (`data-retention.md`): an
  evidence artifact is retained for the audit cycle it supports,
  then ages out. The retention rule for evidence is itself audited
  (deletion of evidence is a meta-audited action, like deletion of
  audit logs).
 - A legal hold suspends evidence deletion for a defined set — the
  same mechanism as audit-log holds.
 ## What Violates Evidence Discipline
 | Violation | Principle |
 |-----------|-----------|
 | Evidence assembled by hand the week before an audit | P6 Evidence is Collected Continuously |
 | An evidence artifact with no provenance (no producer, no timestamp) | P7 Identity is Attributable |
 | An audit packet that requires six weeks of forensics to produce | P6 Evidence is Collected Continuously |
 | An attestation with no signature (provenance asserted, not proven) | P7 Identity is Attributable |
 | Evidence store not queryable between audits | P10 Compliance Posture is Observable |
 | Evidence deleted before its retention period with no meta-audit | P6 Evidence is Collected Continuously |
 | Policy decisions not recorded as evidence | P5 Policy is Evaluated as a Gate |
 | A deployment with no signed attestation of its build provenance | P7 Identity is Attributable |
 ## Relationship to Other Domains
 - `domains/security/supply-chain.md` — signed attestations are the
  supply-chain integrity primitive; evidence.md is the compliance
  consumer of the same artifact.
 - `domains/observability/metrics.md` — posture metrics are a
  continuous evidence stream.
 - `domains/observability/tracing.md` — distributed traces provide
  evidence that spans a request across services.
 - `domains/compliance/audit-logs.md` — audit logs are a primary
  evidence artifact; the append-only guarantee is what makes them
  admissible.
 - `domains/compliance/policy-as-code.md` — policy decisions are
  evidence of enforcement; the policy code itself is evidence of
  the rule.
 - `domains/compliance/data-retention.md` — evidence has a retention
  lifecycle tied to the audit cycle.
@@ -0,0 +1,184 @@
 # Compliance — First Principles
 > Framework-agnostic per D-024. These principles derive from core
 > Security (a subset of C1 Correctness), Observability, and
 > Reversibility. They apply across regulations — NIST CSF, SOC 2,
 > GDPR, CCPA, HIPAA, PCI-DSS, ISO 27001 — without prescribing any
 > regulation-specific implementation. Regulation names appear here
 > only as examples of what the principles support; the principles
 > themselves are engineering rules, not legal controls.
 ## 1. The Principles
 ### P1. Audit Logs are Append-Only
 Audit records are immutable once written. Deletion or mutation of an
 audit record is itself an auditable incident — the tampering is the
 signal, not just the underlying event. An audit log that can be edited
 is not an audit log; it is a draft. Append-only is enforced
 structurally (write-once storage, immutable buckets, hash-chained
 records), not by policy alone. This is the compliance angle on
 `domains/observability/logging.md`: structured logs that cannot be
 rewritten are the substrate of accountability.
 ### P2. Every Significant Action is Logged
 The set of auditable actions is defined a priori, in code, before the
 action ships — not retrofitted after an incident. Authentication
 changes, authorization decisions, data access, configuration changes,
 policy changes, and deletions are all significant. "We forgot to log
 it" is a violation, not an excuse. The auditable-action catalog is
 itself versioned and reviewed. A significant action with no log line
 is a C7 (Observability) defect and a C1 (Correctness) defect: the
 system's behavior is invisible, and accountability is impossible.
 ### P3. Retention is Policy, Not Storage
 Data lifetime is declared and enforced as policy, not left to the
 storage layer's defaults. Deletion at end-of-life is a feature, not a
 failure. Retention rules live as code (lifecycle rules, scheduled
 deletion jobs, tiered storage transitions), they are reviewed, and
 they are auditable. "We kept it because the bucket was cheap" is a
 violation; "we deleted it because the policy said to" is correct.
 Retention is distinct from backup: a backup is a recovery mechanism,
 a retention rule is a deletion mechanism. Keeping them conflated
 produces data that survives both the deletion policy and the
 disaster — which is the opposite of compliance. Cross
 `domains/data/migrations.md` for the schema-lifecycle discipline.
 ### P4. Policy is Code
 Compliance policy is expressed in versioned, reviewable, testable
 code (OPA / Rego, AWS Cedar, HashiCorp Sentinel, Kyverno) — not in
 spreadsheets, prose documents, or tribal knowledge. Policy in a
 spreadsheet is untestable, unreviewable, and undeployable; it is a
 wishlist, not a control. Policy-as-code inherits the disciplines of
 `domains/infrastructure-as-code/P1 Declarative Intent`: declarative
 intent, version control, review before merge, plan before apply. A
 compliance rule that is not executable is a rule that cannot be
 enforced, which is a rule that does not exist.
 ### P5. Policy is Evaluated as a Gate
 Policy violations block before the action, not after the audit.
 Enforcement happens at admission time (kubernetes admission), at
 pipeline time (CI/CD gates), and at provisioning time (IaC plan
 gates) — before the non-compliant state is realized. Detecting a
 violation after it ships is detection, not enforcement. A policy that
 is "logged but not blocked" is a postcard, not a gate. This is the
 compliance angle on C5 (Reversibility): a blocked action is
 reversible by construction; a shipped violation requires remediation,
 which is more expensive than prevention.
 ### P6. Evidence is Collected Continuously
 Evidence of compliance — logs, configs, scans, attestations, policy
 decisions, access reviews — is gathered as a byproduct of operation,
 not assembled manually at audit time. The audit-time scramble is the
 anti-pattern: it is expensive, it is incomplete, and it produces
 evidence that is reconstructed rather than recorded. Continuous
 evidence collection means the audit packet is a query over
 already-collected artifacts, not a forensic reconstruction. This is
 the compliance angle on `domains/observability/tracing.md` for
 distributed evidence and `domains/observability/metrics.md` for
 posture signals.
 ### P7. Identity is Attributable
 Every logged action traces to an authenticated, non-shared principal.
 Shared accounts, generic service identities, and "admin" as an actor
 are violations: an action with no attributable human or workload is
 an action with no accountability. Identity is recorded in the audit
 record at the time of the action, not resolved after the fact. This
 is the compliance angle on `domains/security/authorization.md` and
 `domains/kubernetes/rbac.md`: the audit subject must be the principal
 that acted, not a role that many can assume.
 ### P8. Subject Access is Honored
 Data-subject rights — access, export, deletion, correction — are
 operations with defined contracts and audit trails, not ad-hoc
 tickets. The system can answer "what do we have on this subject,"
 "export it in a portable form," and "delete it and prove the
 deletion" as first-class operations. These are not features bolted on
 at the end; they are contracts the data layer implements from the
 start. A subject-access request that requires a forensics team is a
 correctness defect: the system does not know what it holds. Cross
 `domains/data/schema-design.md` for the data shapes that make
 subject access computable.
 ### P9. Secrets and Sensitive Data are Redacted in Audit
 Audit logs themselves must not leak secrets, credentials, PII, or
 other sensitive data. Redaction is structural — applied at the
 logging boundary, before the record is written — not opportunistic
 scrubbing after the fact. A secret that appears in an audit log is a
 C1 (Correctness) violation (the log is now a secret store) and a
 security violation (`domains/security/secrets.md`). The redaction
 policy is itself auditable: which fields are redacted, by what rule,
 in which log stream. Once a secret is in an append-only log, the
 remediation is expensive — rotate the secret and rewrite the log's
 access scope — so redaction-at-source is the only sound position.
 ### P10. Compliance Posture is Observable
 The system reports its own compliance state: drift from policy, open
 violations, retention status, evidence freshness, policy-evaluation
 counts. Silent non-compliance is the bug. A compliance posture
 metric is a first-class signal (`domains/observability/metrics.md`),
 alertable, and dashboarded. "We didn't know we were non-compliant"
 is not a defense; it is a C7 (Observability) defect. The posture is
 queryable at any time, not only at audit time. This is the compliance
 angle on `domains/infrastructure-as-code/P3 State is Truth`: the
 compliance state is a versioned, queryable truth, not a vibe.
 ## 2. Core Principle Trace
 Each compliance P-rule derives from one or more core C-rules (C1–C8).
 The matrix extension lands in P4 of the v0.3 plan; the traces below
 are authoritative.
 | P-rule | Core | Why |
 |--------|------|-----|
 | P1 Audit Logs are Append-Only | C1, C5 | Correctness of the record; reversibility of tamper detection |
 | P2 Every Significant Action is Logged | C7, C1 | Observability of behavior; correctness of a-priori audit scope |
 | P3 Retention is Policy, Not Storage | C5, C8 | Reversibility of data lifetime; economy of storage as policy |
 | P4 Policy is Code | C6, C2 | Composability of versioned policy; clarity of executable intent |
 | P5 Policy is Evaluated as a Gate | C1, C5 | Correctness of pre-action enforcement; reversibility of blocked actions |
 | P6 Evidence is Collected Continuously | C7, C3 | Observability of compliance state; simplicity of audit-by-query |
 | P7 Identity is Attributable | C1, C7 | Correctness of accountability (security subset); observability of who acted |
 | P8 Subject Access is Honored | C1, C5 | Correctness of the data-subject contract; reversibility of deletion |
 | P9 Secrets and Sensitive Data are Redacted in Audit | C1, C3 | Correctness of not leaking (security subset); simplicity of structural redaction |
 | P10 Compliance Posture is Observable | C7, C1 | Observability of posture; correctness of self-reported state |
 ## 3. What Violates These Principles
 | Violation | Principle Breached |
 |-----------|-------------------|
 | An audit log stored on a mutable filesystem with no write-once protection | P1 Audit Logs are Append-Only |
 | A `DELETE` on an audit record to "clean up a typo" | P1 Audit Logs are Append-Only |
 | An auth change with no audit log line | P2 Every Significant Action is Logged |
 | "We'll add logging after we ship the feature" | P2 Every Significant Action is Logged |
 | Data kept indefinitely because "the bucket is cheap" | P3 Retention is Policy, Not Storage |
 | A retention rule in a spreadsheet, not in code | P4 Policy is Code |
 | A policy that logs violations but does not block the action | P5 Policy is Evaluated as a Gate |
 | Evidence assembled by hand the week before an audit | P6 Evidence is Collected Continuously |
 | A shared `admin` account as the audit actor | P7 Identity is Attributable |
 | A subject-access request that requires a forensics team | P8 Subject Access is Honored |
 | A secret visible in an audit log entry | P9 Secrets and Sensitive Data are Redacted in Audit |
 | No dashboard for compliance posture between audits | P10 Compliance Posture is Observable |
 ## 4. Relationship to Other Domains
 Compliance is the accountability layer that crosses
 `domains/security/` (it audits security actions),
 `domains/observability/` (audit logs are structured logging; posture
 is metrics; evidence is traces), `domains/data/` (retention and
 subject access are data-layer contracts), and
 `domains/infrastructure-as-code/` (policy-as-code parallels
 declarative IaC; compliance state parallels state-as-truth). Cross-
 links are one-directional (per D-026 extended):
 - `domains/security/authorization.md` ← P7 (attributable identity)
 - `domains/security/secrets.md` ← P9 (redaction)
 - `domains/security/supply-chain.md` ← P6 (signed attestations as evidence)
 - `domains/observability/logging.md` ← P1, P2 (audit logs = structured logging)
 - `domains/observability/metrics.md` ← P10 (compliance posture metrics)
 - `domains/observability/tracing.md` ← P6 (evidence from distributed traces)
 - `domains/data/schema-design.md` ← P3, P8 (retention and subject-access shapes)
 - `domains/data/migrations.md` ← P3 (retention migration discipline)
 - `domains/infrastructure-as-code/P1 Declarative Intent` ← P4 (policy-as-code)
 - `domains/infrastructure-as-code/P3 State is Truth` ← P10 (compliance posture truth)
 - `domains/kubernetes/rbac.md` ← P7 (audit subject identity)
 - `domains/devops/ci-cd.md` ← P5 (policy as a pipeline gate)
 - `domains/devops/first-principles.md` ← P4 (policy as configuration-as-code)
@@ -0,0 +1,140 @@
 # Policy as Code — Derived Rules
 > Derives from `domains/compliance/first-principles.md`. Covers P4
 > (Policy is Code) and P5 (Policy is Evaluated as a Gate). Referenced
 > by `audit-logs.md` (policy decisions are audited) and `evidence.md`
 > (policy decisions are evidence). Framework-agnostic per D-024.
 ## Policy is Code (P4 Policy is Code)
 - Compliance policy is expressed in versioned, reviewable, testable
  code — not in spreadsheets, prose documents, or tribal knowledge.
  Policy in a spreadsheet is untestable, unreviewable, and
  undeployable; it is a wishlist, not a control.
 - Policy-as-code inherits the disciplines of
  `domains/infrastructure-as-code/P1 Declarative Intent`: declarative
  intent, version control, review before merge, plan before apply.
  A compliance rule that is not executable is a rule that cannot be
  enforced, which is a rule that does not exist.
 - Policy code is tested like any other code: unit tests for the rule
  logic (given an input, the rule allows or denies as expected),
  integration tests for the gate (the rule fires at the right point
  in the pipeline), and versioning for the policy itself (a policy
  change is a reviewed, merged, deployed change).
 ## Policy is Evaluated as a Gate (P5 Policy is Evaluated as a Gate)
 - Policy violations block **before** the action, not after the
  audit. Enforcement happens at:
  - **Admission time** — a kubernetes admission webhook denies a
    non-compliant resource before it is created
    (`domains/kubernetes/rbac.md`).
  - **Pipeline time** — a CI/CD gate denies a non-compliant change
    before it merges (`domains/devops/ci-cd.md`).
  - **Provisioning time** — an IaC plan gate denies a non-compliant
    resource before `apply` (`domains/infrastructure-as-code/`).
 - A policy that logs violations but does not block the action is a
  postcard, not a gate. Detection is not enforcement. A logged
  violation that the actor could ignore is a P5 violation — the
  policy exists, but the system is not compliant by construction.
 - The gate is the contract. The policy author writes the rule; the
  gate operator wires the rule into the enforcement point; the
  auditor verifies the gate fired. All three are auditable
  (`audit-logs.md`).
 ## Engine Comparison (IDEATE-23)
 | Engine | Policy Language | Evaluation Gate | Ecosystem | Notes |
 |--------|-----------------|-----------------|-----------|-------|
 | **OPA / Rego** | Rego (declarative, set-based, Datalog-inspired) | CI/CD, k8s admission (Gatekeeper), HTTP API, IaC plan (Terraform Sentinel-style), service mesh | Broadest ecosystem; CNCF graduated; library of reusable bundles | General-purpose; the default choice when the gate location varies |
 | **AWS Cedar** | Cedar (declarative, authorization-focused, schema-typed) | k8s admission (via Cedar-agent), application authorization, AVP (Verified Permissions) | AWS-native; tight schema typing; separates policy from entities | Authorization-focused; strong where the policy is "who can do what on which resource" |
 | **HashiCorp Sentinel** | Sentinel (declarative, restricted, policy-focused) | Terraform / TFE plan gate, Nomad, Vault | HashiCorp ecosystem; embedded in Terraform Enterprise / HCP | IaC-plan-gate native; the enforcement point is the `plan` output |
 | **Kyverno** | Kyverno (YAML-declarative, k8s-native, no new DSL) | k8s admission (native), cluster-wide policy reports | Kubernetes-native; no separate language — policy is a CRD | k8s-cluster-gate native; the choice when the gate is admission and the team prefers YAML over a DSL |
 - None is advocated over the others. The choice is (a) where the
  gate fires, (b) the team's tolerance for a new policy language,
  and (c) ecosystem fit. All four satisfy P4/P5 when wired
  correctly.
 - A gate is a gate regardless of engine: the rule is declarative,
  the evaluation is pre-action, and the decision is allow-or-deny.
  The engine difference is language and enforcement-point fit, not
  correctness.
 ## Policy Testing
 - Policy code is unit-tested like any other code. A test asserts
  that a given input produces the expected decision (allow / deny /
  warn). The test is versioned with the policy; a policy change
  with no test change is a red flag.
  ```
  // illustrative Rego policy + test
  // policy: deny containers running as root
  package k8s.admission
  deny[msg] {
    input.kind == "Pod"
    c := input.spec.containers[_]
    not c.securityContext.runAsNonRoot
    msg := sprintf("container %s must set runAsNonRoot", [c.name])
  }
  // test (Rego unit test)
  package k8s.admission
  test_deny_root_container {
    some msg in deny with input as {
      "kind": "Pod",
      "spec": { "containers": [ { "name": "app", "securityContext": {} } ] }
    }
    msg == "container app must set runAsNonRoot"
  }
  ```
 - Integration tests assert the gate fires: a non-compliant resource
  submitted to the admission endpoint is denied; a compliant one is
  allowed. The integration test runs against the real gate, not a
  mock, because the gate wiring is half the contract.
 ## Policy Versioning
 - Policy is versioned in git. A policy change is a reviewed, merged,
  deployed change — the same discipline as application code. A
  policy that is edited in production without review is a P4
  violation: the policy is code, but it is being treated as config.
 - A policy change can break existing workloads (a new deny rule
  blocks a previously-allowed resource). The rollout is staged:
  warn-only mode first (log violations, do not block), then enforce
  mode after the violation count is zero. This is the policy
  analogue of `domains/devops/P5 Progressive Delivery`.
 ## What Violates Policy-as-Code Discipline
 | Violation | Principle |
 |-----------|-----------|
 | A compliance rule in a spreadsheet | P4 Policy is Code |
 | A policy that logs violations but does not block the action | P5 Policy is Evaluated as a Gate |
 | A policy edited in production without review | P4 Policy is Code |
 | A policy with no unit tests for the rule logic | P4 Policy is Code |
 | A gate wired with a mock instead of the real engine | P5 Policy is Evaluated as a Gate |
 | A new deny rule enforced without a warn-only rollout | P5 Policy is Evaluated as a Gate |
 | A policy in prose ("the team should not use root containers") | P4 Policy is Code |
 | A policy decision with no audit record | P5 Policy is Evaluated as a Gate |
 ## Relationship to Other Domains
 - `domains/infrastructure-as-code/first-principles.md` — policy-as-
  code inherits declarative intent, versioning, and plan-before-
  apply from IaC.
 - `domains/kubernetes/rbac.md` — k8s admission is a primary
  enforcement gate; Kyverno and OPA Gatekeeper wire into it.
 - `domains/devops/ci-cd.md` — CI/CD is a pipeline-time enforcement
  gate; a policy step blocks a non-compliant change before merge.
 - `domains/compliance/audit-logs.md` — every policy decision (allow
  / deny) is an audited significant action.
 - `domains/compliance/evidence.md` — policy decisions and the
  policy code itself are evidence of enforcement posture.
 - `domains/security/authorization.md` — Cedar's authorization-
  focused policy overlaps with authz; the split is that authz is
  the runtime decision, policy-as-code is the reviewed rule that
  drives it.
@@ -0,0 +1,227 @@
 # CDN — Derived Rules
 > Derives from `domains/edge/first-principles.md`. Applies P1
 > (Proximity is the Design Driver) and P6 (Cache Invalidation is
 > Explicit) primarily, with P5 (idempotent cache fill), P8
 > (geographic distribution), and P9 (identity at the edge). For the
 > edge-cache-vs-origin decision, see the decision matrix below.
 > Cross-links `domains/performance/frontend` for generic caching,
 > `domains/security/input-validation` for cache poisoning, and
 > `domains/observability/metrics` for cache-hit ratio.
 ## What a CDN Is (P1 Proximity is the Design Driver)
 - A content delivery network is a fleet of PoPs (points of presence)
  placed near users. The PoP serves cached content; the origin is
  the authoritative source. The CDN's whole purpose is P1: compute
  (the cache) is placed near the user so the round trip to the origin
  does not bound latency. Latency is a correctness constraint at the
  edge (C1), not a performance preference.
 - The CDN is the canonical edge-cache architecture (Akamai,
  Cloudflare, Fastly): PoPs near users, origin shielding, cache-key
  normalization, purge APIs. Atelier derives the
  placement/invalidation principles, not the vendor config.
 - The boundary with `domains/performance/frontend` is per D-061:
  performance owns *generic* caching and optimization (cache what is
  expensive, stable, read often — `performance/P5 Caching with
  Intent`); edge owns the *geographic, partition-aware* placement and
  invalidation angle. A CDN is an edge concern because its defining
  trait is geographic distribution (P8) and partition-aware
  invalidation (P6), not measurement.
 ## Cache Key Design (P6 Cache Invalidation is Explicit)
 - The cache key is the contract between the URL and the cached
  representation. A key that varies on the wrong dimensions serves
  the wrong content; a key that varies on too many dimensions
  collapses the hit ratio. Key design *is* the invalidation
  surface: a key that includes a content hash or version segment
  makes invalidation explicit; a key that ignores `Vary` headers
  serves stale variants.
 - Normalize the key: lower-case the host, strip default ports,
  sort query parameters, ignore tracking parameters. A
  non-normalized key is a cache-poisoning vector (see
  `domains/security/input-validation`) and a hit-ratio destroyer
  (see `domains/observability/metrics`).
 - A cache with no explicit key strategy is a TTL-less cache under
  partition (P6 violation): staleness is silent and unbounded.
 ```http
 # Cache key derivation: vary on what changes content, ignore what
 # does not. The key is the tuple (host, normalized-path, sorted-
 # query, Vary-headers); the cache entry is the representation + TTL.
 Cache-Key: example.com /api/v1/products?sort=price&region=us Vary:Accept-Encoding
 Cache-Control: public, max-age=60, s-maxage=600, stale-while-revalidate=300
 Vary: Accept-Encoding
 ```
 - `max-age` bounds the browser cache; `s-maxage` bounds the CDN
  PoP; `stale-while-revalidate` allows serving stale while
  refetching. Each is an explicit invalidation strategy (P6).
 ## TTL vs Explicit Invalidation (P6, C3 Simplicity)
 - **TTL-based invalidation** (`max-age`, `s-maxage`): the cache entry
  expires after a duration. Simple, no origin contact required to
  invalidate, but bounded staleness is the contract — the entry may
  be stale up to TTL. Fits content where eventual consistency is
  acceptable (asset fingerprints, lists, derived images).
 - **Explicit invalidation** (purge, surrogate keys): the operator
  signals the cache to drop entries. Tighter staleness bounds, but
  requires the origin or operator to know which entries to purge.
  Fits content where staleness is a correctness defect (price
  updates, availability, breaking news).
 - A TTL-less cache with no explicit invalidation is a P6 violation:
  stale-forever under partition. Every cache must have one or the
  other (or both), and the choice is documented per content type.
 ## Cache-Hit / Miss / Origin-Fetch (P5, P6)
 - **Hit**: the PoP serves from cache. Latency is PoP-local (P1).
 - **Miss**: the PoP has no entry; it fetches from the origin (or an
  origin-shield PoP). The fetch must be idempotent (P5) — a retried
  miss must not corrupt the cache or double-write side effects.
 - **Revalidate**: the PoP holds a stale entry and asks the origin
  (`If-None-Match`, `If-Modified-Since`); a 304 refreshes the TTL
  without re-fetching the body. Revalidation is the bandwidth-economical
  middle ground (C8).
 ```http
 # Conditional revalidation — the PoP asks the origin "is this still
 # current?" The 304 response refreshes the TTL without a body.
 GET /api/v1/products HTTP/1.1
 Host: example.com
 If-None-Match: "etag-7a3f"
 HTTP/1.1 304 Not Modified
 ETag: "etag-7a3f"
 Cache-Control: s-maxage=600
 ```
 - A cache-hit ratio that is not measured is a gate on noise — see
  `domains/observability/metrics` for the SLI/SLO discipline that
  makes the hit ratio a meaningful signal. A CDN with no hit-ratio
  metric is operating blind (P10 analog).
 ## Origin Shielding (P1, P8, C8 Economy)
 - Origin shielding routes all origin fetches through a single
  shield PoP (or shield region). The shield absorbs the
  thundering-herd: 10 000 PoPs missing the same URL fetch the origin
  once, not 10 000 times. This is C8 Economy (origin bandwidth is
  bounded) and P1 (the shield is itself a proximity layer for the
  origin).
 - Shielding is a geographic decision (P8): the shield sits in a
  region close to the origin, not close to the user. The shield is
  the inner ring of the CDN; the user-facing PoPs are the outer ring.
 - A CDN without origin shielding under a stampede will overload the
  origin; a shield that is itself partitioned from the origin must
  degrade gracefully (P7) — serve stale per `stale-while-revalidate`
  rather than 500.
 ## Purge Strategies (P6, C3 Simplicity)
 | Strategy | Granularity | Latency to Invalidate | Cost | Best for |
 |----------|-------------|-----------------------|------|----------|
 | URL purge | One URL | Seconds | Low (one entry) | Surgical fixes, single-page corrections |
 | Soft purge | One URL (mark stale, serve while refetch) | Seconds | Low | High-traffic URLs where a hard purge causes a stampede |
 | Surrogate-key purge | A tag set (e.g., `product:123`, `category:shoes`) | Seconds | Medium (key indexing) | Related-content invalidation (a product update purges all its category pages) |
 | Wildcard purge | A path prefix or pattern | Seconds to minutes | High (scan) | Site-wide template changes |
 | All-cache purge | Everything | Seconds | Very high (origin stampede) | Disaster recovery only; never the steady-state invalidation path |
 - Surrogate-key purge (Fastly, Akamai) is the highest-value
  strategy: tag cache entries with content keys, then purge by tag.
  This is explicit invalidation at scale (P6) without the origin
  stampede of an all-cache purge.
 - An all-cache purge as the steady-state invalidation path is a P6
  violation dressed as a feature — it pushes the origin load back to
  100% miss, defeating the CDN's purpose (P1).
 ## Cache Poisoning Prevention (P9, cross-link security/input-validation)
 - A cache poisoned by a crafted request (a URL with a malicious
  header that gets cached and served to others) is a correctness
  defect (C1) and a security breach (P9 — the edge node is
  exploited). Prevent poisoning by:
  - Normalizing the cache key (strip untrusted query parameters,
    ignore unknown headers, lower-case the host).
  - Validating `Vary` against an allow-list; never `Vary: *` on a
    shared cache (poisonable via header injection).
  - Treating uncacheable responses (`Set-Cookie`,
    `Cache-Control: private`) as never-stored.
 - See `domains/security/input-validation` for the general
  input-validation discipline the cache key must follow. The cache
  key is a validation surface; a non-validated key is an attack
  surface.
 ## Multi-CDN Routing (P8 Geographic Distribution)
 - A multi-CDN strategy routes each request to the best PoP across
  providers (Akamai + Cloudflare + Fastly). Routing is
  location-aware (P8): latency, cost, and availability vary by
  region and provider. The DNS layer (or a client-side router)
  selects the CDN per request.
 - Multi-CDN is a P8 decision, not a vendor-management decision:
  geographic distribution is the first-class constraint. A
  single-CDN deployment routes everything to one provider's PoPs;
  a multi-CDN deployment routes by region, latency, and cost.
 - Invalidation across multiple CDNs is harder (P6): each provider
  has its own purge API and surrogate-key scheme. A multi-CDN purge
  must fan out to all providers; a purge that reaches only one CDN
  leaves the others stale. Track purge completion per provider —
  see `domains/observability/metrics` for the per-CDN hit-ratio and
  purge-latency signals.
 ```http
 # A CDN config example: cache-control headers + a purge rule.
 # Origin response: declare the cache contract (P6).
 HTTP/1.1 200 OK
 Cache-Control: public, max-age=60, s-maxage=600, stale-while-revalidate=300
 Surrogate-Key: product:123 category:shoes
 ETag: "etag-7a3f"
 Vary: Accept-Encoding
 # Purge rule (Fastly-style surrogate-key): when product 123
 # updates, purge every cache entry tagged product:123 OR
 # category:shoes. Explicit, bounded, no origin stampede (P6).
 POST /service/svc1/purge
 Surrogate-Key: product:123 category:shoes
 # Returns: {"status": "ok", "id": "purge-abc"} — poll the purge
 # status to confirm completion across all PoPs (P8, P10).
 ```
 ## Edge-Cache vs Origin — Decision Matrix (D-069)
 | Strategy | When | Latency | Origin Load | Correctness Risk |
 |----------|------|---------|-------------|------------------|
 | Serve from PoP (cache hit) | The PoP holds a fresh entry (within TTL or revalidated) | Lowest (PoP-local, P1) | None | Low — bounded by TTL staleness (P6) |
 | Serve stale while revalidate | The PoP holds a stale entry and `stale-while-revalidate` is set | Low (stale served immediately, refetch in background) | Background refetch (1 per entry) | Medium — stale served up to the revalidate window; acceptable for eventually-consistent content |
 | Fetch fresh from origin (miss) | The PoP has no entry, or the content is non-cacheable | High (origin round trip) | Full fetch per miss | Low — fresh by construction; the miss is the correctness floor |
 | Origin-shield fetch | Multiple PoPs miss the same URL; the shield collapses the herd | Medium (PoP → shield → origin) | Bounded to one origin fetch per shield (C8) | Low — shield is the inner ring; staleness bounded by shield TTL |
 | Purge and serve fresh | Explicit invalidation received (surrogate-key or URL purge) | Medium (purge propagates, then fresh fetch) | Full fetch post-purge | Lowest — explicit invalidation is the tightest staleness bound (P6) |
 | Serve from origin directly (bypass cache) | Content is non-cacheable (personalized, real-time) | Highest (every request hits origin) | Full fetch per request | Lowest for correctness, highest for origin load — use sparingly |
 - The default is **serve from PoP** when fresh, **fetch fresh from
  origin** on miss with **origin-shield** to bound origin load, and
  **purge and serve fresh** when explicit invalidation is required.
  Bypass-the-cache is for non-cacheable content only — bypassing for
  cacheable content is a P1 violation (you have defeated the CDN).
 - The correctness risk column is bounded by the invalidation
  strategy (P6): every row except "bypass" carries staleness risk
  that is bounded by TTL or explicit purge. A row with no
  invalidation strategy is a P6 violation.
 ## What Violates CDN Discipline
 | Violation | Principle |
 |-----------|-----------|
 | TTL-less edge cache under partition (stale-forever, no explicit invalidation) | P6 Cache Invalidation is Explicit |
 | Cache key that varies on untrusted query parameters (poisonable) | P6, P9 (`domains/security/input-validation`) |
 | All-cache purge as the steady-state invalidation path (origin stampede) | P6, C8 Economy |
 | Non-normalized cache key (case-sensitive host, unsorted query) | P6, `domains/security/input-validation` |
 | Bypass-the-cache for cacheable content | P1 Proximity is the Design Driver (defeats the CDN) |
 | Multi-CDN with no per-CDN purge completion tracking | P8, P10 (stale cache invisible to the operator) |
 | Origin fetch that is not idempotent under retry | P5 Edge Operations are Idempotent |
 | Cache-hit ratio not measured | P10, `domains/observability/metrics` |
 | Single-CDN deployed where geographic distribution requires multi-CDN | P8 Geographic Distribution |
 | Shield PoP that 500s instead of serving stale under partition | P7 Partial Degradation is Engineered |
@@ -0,0 +1,218 @@
 # Edge — First Principles
 ## 1. The Principles
 ### P1. Proximity is the Design Driver
 Compute, storage, and data are placed near the user or the data
 source. At the edge, latency is a correctness constraint (C1), not a
 performance preference — a late answer is a wrong answer when the
 round trip to a central region exceeds the user's or device's
 tolerance. This is the geographic expression of `C4 Locality`:
 performance's locality is algorithmic (data near compute); edge's
 locality is geographic (compute near user/data source). Placement is
 a design decision, not an accident of deployment, and it is
 constrained by `P8 Geographic Distribution`. The proximity angle is
 the distinguishing trait of the edge domain per D-061: this is what
 separates edge from `domains/performance/` (which owns *generic*
 measurement and optimization, not placement).
 ### P2. Offline is a First-Class State
 The system continues to operate when disconnected from the center.
 Partition is the norm, not the exception; reconciliation happens on
 reconnect, never assumed to be instant. An app that crashes on
 disconnect has no offline state and is unengineered. Offline
 operation derives from `C5 Reversibility` — the disconnected state
 is reversible back to consistency via reconciliation — and `C1
 Correctness`, because correctness under partition is the contract,
 not eventual correctness as a hedge. This is the foundation for
 `domains/edge/offline-first.md` and the precondition for the
 bounded-conflict discipline of `P4`.
 ### P3. Resources are Constrained and Declared
 Edge nodes — IoT sensors, gateways, point-of-sale devices, CDN PoP
 caches, 5G MEC nodes — have bounded CPU, memory, power, and
 bandwidth. Constraints are declared per node class, never assumed
 infinite. An undeclared budget is a defect: unbounded growth is a
 bug, and a constrained device with no budget will OOM or exhaust
 power. This derives from `C8 Economy` (use no more than the task
 requires) and `C1 Correctness` (a node that exceeds its bounds has
 failed). This is the edge-specific angle on `domains/performance/P4
 Resource Bounds` — performance owns the generic principle; edge owns
 the constrained-device reality. See `domains/edge/iot.md` for the
 per-device-class application.
 ### P4. Sync Conflicts are Bounded, Not Infinite
 Divergent state across partitioned nodes converges. Oscillation and
 infinite sync loops are correctness failures, not eventual
 consistency. A merge that never terminates is a livelock; a CRDT
 without merge semantics or an LWW without a monotonic clock can
 oscillate forever. This derives from `C1 Correctness` (convergence is
 a correctness contract) and `C5 Reversibility` (divergent state is
 reversible back to convergence). The bound may be eventual (CRDTs) or
 arbitrated (LWW with vector clocks), but it must exist. This is the
 foundation for `domains/edge/sync.md` and the rule the
 `edge-sync-loop` chaos anti-pattern breaches.
 ### P5. Edge Operations are Idempotent
 Sync, cache fill, and device commands are retried by nature — the
 network is partition-prone and the operation will be re-attempted.
 Idempotency keys (or deterministic operations) make retries safe. A
 non-idempotent edge write retried with side effects doubles the
 effect; a non-idempotent cache fill under retry corrupts the cache.
 This derives from `C1 Correctness`: correctness under retry is the
 contract, not a nice-to-have. This parallels `domains/messaging/P3
 Consumers are Idempotent` (cross-process delivery) and is the edge's
 device-and-cache-flavored analog — see `domains/edge/iot.md` for
 device command idempotency and `domains/edge/cdn.md` for cache-fill
 idempotency.
 ### P6. Cache Invalidation is Explicit
 Edge caches carry a defined invalidation or TTL strategy. A
 stale-forever cache under partition is a silent correctness defect;
 a TTL-less cache with no explicit invalidation is a bug, not a
 feature. This derives from `C1 Correctness` (cached state must be
 correct) and `C3 Simplicity` (a defined invalidation strategy is
 simpler and clearer than ad-hoc staleness). This is distinct from
 `domains/performance/P5 Caching with Intent`, which owns *generic*
 caching and optimization; edge owns the *geographic,
 partition-aware* invalidation angle — when a PoP is partitioned from
 the origin, the invalidation strategy is the correctness mechanism.
 See `domains/edge/cdn.md` for purge strategies (URL vs soft vs
 surrogate-key) and the edge-cache-vs-origin decision matrix.
 ### P7. Partial Degradation is Engineered
 The system degrades gracefully when an edge node or link fails. A
 partial service is a designed state with a defined contract, not a
 crash. One node's failure must not collapse the whole fleet; the
 degraded mode is documented, observable, and recoverable. This
 derives from `C1 Correctness` (the degraded contract is a
 correctness bound) and `C5 Reversibility` (recovery from degradation
 is reversible by construction). A crash-on-node-failure system has
 no degradation contract — it has an all-or-nothing failure mode that
 violates the fleet assumption. See `domains/edge/iot.md` for
 device-drop degradation and `domains/edge/offline-first.md` for
 partition degradation.
 ### P8. Geographic Distribution is a First-Class Constraint
 The fleet is geo-distributed; routing, fan-out, and data placement
 are location-aware decisions, not accidents of deployment. The
 system is many nodes across many locations, not a single deployment.
 Data residency, regional latency, and PoP selection are engineered,
 not discovered in production. This derives from `C4 Locality` (the
 placement of data and compute is a locality decision) and `C6
 Composability` (the fleet composes from location-aware parts, each
 with its own contract). This is the structural companion to `P1
 Proximity`: P1 says *where* compute should be (near the user); P8
 says the *distribution* of compute across geographies is a
 first-class constraint. See `domains/edge/cdn.md` for multi-CDN
 routing.
 ### P9. Identity is Constrained at the Edge
 Edge devices and nodes hold scoped, minimal credentials. No edge
 node is a cluster-admin-equivalent; device identity is per-device,
 not shared. One compromise must not equal a fleet compromise. This
 derives from `C1 Correctness` (security is a subset of correctness —
 an exploitable edge node does not do what it was supposed to do) and
 `C8 Economy` of trust (the credential scope is minimal for the task).
 A shared edge-device credential is the edge analog of a
 cluster-admin GitOps robot — blast radius is unbounded. See
 `domains/security/secrets.md` for the general secret-hygiene
 principles and `domains/edge/iot.md` for device provisioning.
 ### P10. Edge Observability Survives Partition
 Telemetry is local-first: buffered on the node and forwarded on
 reconnect. Partition does not blind the operator. A fire-and-forget
 telemetry pipeline loses data when the link drops; a local-first
 buffer survives. This derives from `C7 Observability` (the fleet's
 behavior is visible to the operator) and `C5 Reversibility` (the
 buffered telemetry is reversible back to visibility on reconnect).
 This is distinct from `domains/observability/P1 Structured by
 Default`, which owns *generic* structured telemetry; edge owns the
 *partition-survivable, local-first* angle. See
 `domains/observability/metrics.md` and
 `domains/observability/logging.md` for the generic structured-
 telemetry foundations edge builds on.
 ## 2. Core Principle Trace
 Each edge P-rule derives from one or more core C-rules (C1–C8). The
 matrix extension lands in P4 of the v0.4 plan; the traces below are
 authoritative. Edge is a broad-derivation domain touching 7 of 8
 core principles (C1, C3, C4, C5, C6, C7, C8); C2 (Clarity) is not a
 primary derivation — edge clarity is indirect (a cache with explicit
 invalidation is clearer than one without, but the primary trace is
 C1/C3).
 | P-rule | Core | Why |
 |--------|------|-----|
 | P1 Proximity is the Design Driver | C4, C1 | Locality of compute near user/data; correctness via latency |
 | P2 Offline is a First-Class State | C1, C5 | Correctness under partition; reversibility of reconciliation |
 | P3 Resources are Constrained and Declared | C8, C1 | Economy of constrained nodes; correctness of declared bounds |
 | P4 Sync Conflicts are Bounded, Not Infinite | C1, C5 | Correctness of convergence; reversibility of divergent state |
 | P5 Edge Operations are Idempotent | C1 | Correctness under retry |
 | P6 Cache Invalidation is Explicit | C1, C3 | Correctness of cached state; simplicity of defined invalidation |
 | P7 Partial Degradation is Engineered | C1, C5 | Correctness of degraded modes; reversibility of recovery |
 | P8 Geographic Distribution is a First-Class Constraint | C4, C6 | Locality of placement; composability of the fleet |
 | P9 Identity is Constrained at the Edge | C1, C8 | Correctness via security; economy of trust |
 | P10 Edge Observability Survives Partition | C7, C5 | Observability of the fleet; reversibility of buffered telemetry |
 ## 3. What Violates These Principles
 | Violation | Principle Breached |
 |-----------|-------------------|
 | Central-region-only deployment for a latency-bound workload | P1 Proximity is the Design Driver |
 | App that crashes on disconnect (no offline state) | P2 Offline is a First-Class State |
 | Undeclared edge-node resource budget (assumes infinite CPU/memory) | P3 Resources are Constrained and Declared |
 | Sync loop that oscillates forever (CRDT without merge-semantics, LWW without monotonic clock) | P4 Sync Conflicts are Bounded, Not Infinite |
 | Non-idempotent edge write (cache-fill or device command retried with side effects) | P5 Edge Operations are Idempotent |
 | TTL-less edge cache under partition (stale-forever, no explicit invalidation) | P6 Cache Invalidation is Explicit |
 | Crash-on-node-failure (no partial-degradation contract) | P7 Partial Degradation is Engineered |
 | Random geographic placement (no location-aware routing) | P8 Geographic Distribution is a First-Class Constraint |
 | Shared edge-device credential (one key for the whole fleet) | P9 Identity is Constrained at the Edge |
 | Fire-and-forget telemetry (no on-node buffer; data lost on partition) | P10 Edge Observability Survives Partition |
 | Blocking call on a constrained IoT device with no timeout | P3, P5 (blocks the node; retry unsafe without idempotency) |
 | Multi-CDN routing with no PoP-selection logic (latency uncontrolled) | P8, P1 (placement not a design decision) |
 ## 4. Relationship to Other Domains
 Edge computing is the engineering discipline of placing compute,
 storage, and data **near the source of generation or consumption**
 rather than in a centralized cloud. The distinguishing constraints are
 latency-bound operation, resource-constrained nodes,
 geo-distribution as a fleet, and partition-prone operation. Edge
 overlaps three existing domains by *subject* but not by *angle*: per
 D-061, edge owns the proximity/location/constraint/disconnection
 concerns that only arise at the network edge. The C4 Locality
 emphasis is the discriminator: performance's locality is algorithmic
 (data near compute); edge's locality is geographic (compute near
 user/data source). Cross-links are one-directional outward (per
 D-026 extended); no back-link edits to v0.1/v0.2/v0.3 content.
 - `domains/performance/frontend` ← P6 (edge owns geographic,
  partition-aware cache invalidation; performance owns *generic*
  caching and measurement — D-061 boundary)
 - `domains/performance/P4 Resource Bounds` ← P3 (edge owns
  constrained-device reality; performance owns the generic
  unbounded-growth-is-a-bug principle)
 - `domains/observability/metrics` ← P10 (cache-hit ratio, edge
  telemetry aggregation; edge owns the local-first angle, observability
  owns generic structured metrics)
 - `domains/observability/logging` ← P10 (local-first logging buffered
  on-node and forwarded on reconnect)
 - `domains/concurrency/patterns` ← P5 (the offline write-queue is
  the cross-partition analog of the in-process bounded buffer —
  concurrency owns in-process; edge owns partition-survivable)
 - `domains/security/secrets` ← P9 (device credentials are scoped,
  per-device, never shared — edge owns the constrained-identity
  angle; security owns the general secret hygiene)
 - `domains/security/input-validation` ← P6 (cache poisoning
  prevention — edge cache keys are a validation surface)
 - `domains/data/migrations` ← P4 (schema migration under sync must
  reconcile across partitioned nodes; data owns the generic migration
  discipline, edge owns the partitioned-reconcile angle)
 > Note: cross-links to `domains/messaging/` (e.g., MQTT QoS parallels
 > for delivery semantics) are intentionally omitted here — the
 > messaging domain is authored in P2. The intra-v0.4 edge↔messaging
 > links are added in P5 (ATELIER-114 per IDEATE-40) once both
 > domains exist; the dangling link is acceptable per D-053.
@@ -0,0 +1,258 @@
 # IoT — Derived Rules
 > Derives from `domains/edge/first-principles.md`. Applies P3
 > (Resources are Constrained and Declared) primarily, with P5
 > (command idempotency), P7 (partial degradation when devices drop),
 > P9 (device identity and provisioning), and P10 (telemetry from
 > devices). Cross-links `domains/security/secrets` for device
 > credentials and `domains/messaging/queues` for the MQTT QoS
 > parallels to delivery semantics.
 ## What IoT at the Edge Is (P3 Resources are Constrained and Declared)
 - IoT at the edge is the engineering discipline of operating
  constrained devices — sensors, actuators, gateways, microcontrollers
  — as first-class participants in a distributed system. The
  distinguishing constraint is per-device resource bounds (P3): a
  battery-powered sensor has kilobytes of RAM, a constrained
  protocol, and a multi-year sleep budget. These constraints are
  declared per device class, never assumed infinite.
 - The boundary is per D-061: edge owns the constrained-device
  reality; performance owns the generic unbounded-growth-is-a-bug
  principle (`performance/P4 Resource Bounds`); concurrency owns
  in-process primitives. IoT is an edge concern because its defining
  traits are constrained resources (P3), geographic distribution as
  a fleet (P8), partition-prone operation (P2), and device-scoped
  identity (P9) — concerns that only arise at the network edge.
 - See `domains/edge/offline-first.md` for the partition-survival
  discipline that constrained devices depend on, and
  `domains/edge/sync.md` for the reconciliation of device state
  across partitions.
 ## Device Resource Classes (P3, C8 Economy)
 - A device resource class declares the bounds for a class of
  devices: CPU (MHz, cores), memory (KB/MB), power (battery mAh,
  duty-cycle budget), bandwidth (bytes/sec, latency budget), and
  storage (KB/MB). Every device in the fleet is assigned to a class;
  every operation is budgeted against its class.
 - An undeclared budget is a defect (P3 violation): a sensor that
  sends telemetry every second without a duty-cycle budget exhausts
  its battery in days, not years. The budget is the correctness
  bound (C1) and the economy bound (C8).
 - A device class implies a protocol choice: a class-0 device
  (constrained sensor, KB RAM) speaks CoAP; a class-1 device
  (gateway, MB RAM) speaks MQTT; a class-2 device (edge compute
  node, GB RAM) speaks HTTP. The protocol follows the constraint,
  not the reverse.
 | Class | RAM | Power | Protocol | Typical role |
 |-------|-----|-------|----------|--------------|
 | 0 (constrained sensor) | < 10 KB | Battery, multi-year | CoAP, LoRaWAN | Telemetry only, no inbound commands |
 | 1 (actuator, gateway) | 10 KB – 1 MB | Battery or wired, weeks-months | MQTT, CoAP | Telemetry + commands, queue-and-forward |
 | 2 (edge compute) | > 1 MB | Wired, continuous | HTTP, MQTT | Local aggregation, gateway, edge inference |
 ## Constrained Protocols — MQTT and CoAP (P3, P5)
 - **MQTT** is the canonical pub/sub protocol for constrained devices.
  It is lightweight (2-byte header), broker-backed, and provides QoS
  levels (0, 1, 2) that map to delivery semantics. MQTT is the
  cross-process analog of message-queue delivery — see
  `domains/messaging/queues` for the general queue/delivery-semantics
  discipline; the cross-link is one-directional outward (edge →
  messaging) per D-062 and D-026 extended.
 - **CoAP** is the REST analog for constrained devices: UDP-based,
  low-overhead, with confirmable (CON) and non-confirmable (NON)
  message types. CoAP fits class-0 devices where TCP is too heavy.
 - A blocking synchronous call on a constrained device with no
  timeout is the `blocking-call-on-constrained-device` chaos
  anti-pattern: it blocks the node, has no timeout (= hang), and
  retries are unsafe without idempotency (P3 + P5 breach). Every
  device operation must be async with a timeout, and every retried
  operation must be idempotent.
 ```json
 // MQTT publish/subscribe payload with QoS levels (P5 idempotency,
 // P3 constrained protocol).
 // QoS 0 — at-most-once: fire-and-forget, no ack. For telemetry
 // where a dropped sample is acceptable (P3 economy of the
 // constrained link).
 {
  "topic": "devices/sensor-7/temperature",
  "qos": 0,
  "payload": {
    "device": "sensor-7",
    "ts": 1700000000,
    "value": 21.4,
    "unit": "C"
  }
 }
 // QoS 1 — at-least-once: acked, may duplicate. The consumer must
 // be idempotent (P5) — dedup by (device, ts) or an idempotency key.
 {
  "topic": "devices/actuator-3/command",
  "qos": 1,
  "payload": {
    "device": "actuator-3",
    "idempotencyKey": "cmd-1700000000-1",
    "command": "set-point",
    "value": 22.0
  }
 }
 // QoS 2 — exactly-once: four-step handshake, no duplication. The
 // heaviest QoS; use only where duplicates are intolerable AND the
 // device has the budget for the handshake (class-1+ only, P3).
 {
  "topic": "devices/actuator-3/irreversible-command",
  "qos": 2,
  "payload": {
    "device": "actuator-3",
    "idempotencyKey": "cmd-1700000000-2",
    "command": "calibrate"
  }
 }
 ```
 - The QoS choice is a P5 (idempotency) and P3 (resource) decision:
  QoS 0 is cheapest (no ack) but lossy; QoS 1 requires consumer
  idempotency (P5); QoS 2 is exactly-once but costs a four-step
  handshake on a constrained link. The default for telemetry is QoS
  0 or 1; the default for commands is QoS 1 with an idempotency key
  (P5); QoS 2 is reserved for irreversible commands where the
  device budget permits.
 ## Device Identity and Provisioning (P9 Identity is Constrained at the Edge)
 - Every device holds a per-device identity: a unique device ID and a
  scoped credential (X.509 certificate, API token, or rotating
  key). No shared fleet credential — one compromise must not equal
  a fleet compromise (P9). The credential scope is minimal: a
  device can publish to `devices/<its-id>/+` and subscribe to
  `devices/<its-id>/commands`, nothing else.
 - Provisioning is the act of assigning a device identity at
  enrollment time. The provisioning manifest declares the device,
  its class, its allowed topics, and its credential. The manifest is
  the P9 contract — a device operating outside its manifest scope is
  a violation.
 - A device that is provisioned with a shared fleet key (the
  `shared-edge-device-credential` anti-pattern) is a P9 violation:
  blast radius is unbounded. See `domains/security/secrets` for the
  general secret-hygiene discipline (per-identity credentials,
  rotation, minimal scope) that device provisioning builds on.
 ```yaml
 # Device provisioning manifest (P9 per-device identity + scoped
 # credentials). The manifest is the contract; the device operates
 # only within its declared scope.
 device:
  id: sensor-7
  class: 0                    # P3 resource class
  model: temp-sensor-v2
  firmware: 1.4.2
 identity:
  cert: "sha256-of-device-cert"
  credentialScope:
    publish:
      - "devices/sensor-7/temperature"
      - "devices/sensor-7/status"
    subscribe:
      - "devices/sensor-7/commands"
    # No wildcard, no fleet-wide topics (P9).
 provisioning:
  enrolledAt: 2024-01-15T00:00:00Z
  rotatesEvery: 90d
  # Per-device credential; never shared (P9, domains/security/secrets).
 ```
 ## Telemetry from Devices (P10 Edge Observability Survives Partition)
 - Device telemetry is local-first (P10): the device buffers telemetry
  on-node and forwards on reconnect. A fire-and-forget telemetry
  pipeline loses data when the link drops; a buffered pipeline
  survives. The buffer is bounded by the device class (P3): a
  class-0 sensor buffers minutes of telemetry, not hours.
 - Telemetry is observable in aggregate: the operator sees the fleet's
  behavior, not just per-device. A device that has not reported in
  its expected interval is itself a signal (a dead device, a
  partitioned device, a drained battery). See
  `domains/observability/metrics` for the generic structured-metrics
  discipline; edge owns the partition-survivable, local-first angle.
 - Telemetry must not be a secrets channel (P9 analog, see
  `domains/observability/P6 No Secrets in Observability`): device
  credentials, PII, and personally-identifying location must not
  enter telemetry payloads.
 ## Command Idempotency (P5 Edge Operations are Idempotent)
 - Device commands are retried by nature (the network is
  partition-prone). Every command carries an idempotency key so a
  retried command does not double-apply (P5). A `set-point` command
  retried with the same idempotency key sets the point once, not
  twice; an `open-valve` command retried is safe because the valve
  is already open.
 - Irreversible commands (a calibration burn-in, a firmware flash)
  require stronger idempotency: the device tracks applied
  idempotency keys and refuses re-application. A retried irreversible
  command without idempotency tracking double-applies the effect
  (P5 violation, possibly a physical-side-effect bug).
 - The idempotency key is per-command, not per-device. A device that
  dedups by device ID alone will drop distinct commands issued in
  the same window. Use `(device, command-id, ts-window)` or a
  UUID per command.
 ## Partial Degradation When Devices Drop (P7 Partial Degradation is Engineered)
 - A fleet degrades when devices drop (battery exhaustion, partition,
  hardware failure). The system must continue to operate with the
  remaining devices; a whole-system crash on one device's failure is
  a P7 violation. The degraded mode is documented: which
  aggregations are valid with N-1 devices, which alerts fire, which
  fallbacks engage.
 - A device that drops is not an incident by itself — fleets expect
  churn. The operator-facing signal is the *aggregate* health (X%
  of devices reporting, Y% partitioned for >Z minutes), not the
  per-device drop. Per-device drop alerts are noise; aggregate
  degradation alerts are signal (see `domains/observability/metrics`).
 - A command to a dropped device must time out (P5 — idempotent
  retry) and degrade (P7 — the fleet continues without that
  device). A command that blocks forever waiting for a dropped
  device is the `blocking-call-on-constrained-device` chaos
  anti-pattern (P3 + P5 breach).
 ## Cross-Link to Messaging (P5, cross-link messaging/queues)
 - MQTT QoS 0/1/2 maps to at-most-once / at-least-once / exactly-once
  delivery semantics — the same three-way tradeoff documented in
  `domains/messaging/queues`. The cross-link is one-directional
  outward (edge → messaging) per D-026 extended: edge owns the
  constrained-device protocol angle; messaging owns the generic
  cross-process delivery-semantics angle.
 - This link dangles until P2 (the messaging domain is authored in
  P2); it is verified bidirectional in P5 (ATELIER-114 per
  IDEATE-40). Acceptable per D-053 (vertical-slice integrity — P1
  ships the edge domain self-consistent; the messaging cross-link
  resolves by the P6 ship).
 - The parallel: a constrained device's QoS 1 publish is the
  device-flavored instance of an at-least-once queue delivery — the
  consumer (the broker or the downstream service) must be
  idempotent (P5 here, `messaging/P3 Consumers are Idempotent`
  there). The idempotency discipline is the same; the protocol and
  failure model differ (constrained-device link vs broker-backed
  network).
 ## What Violates IoT-at-the-Edge Discipline
 | Violation | Principle |
 |-----------|-----------|
 | Undeclared device resource budget (assumes infinite battery/RAM) | P3 Resources are Constrained and Declared |
 | Shared fleet credential (one key for all devices) | P9 Identity is Constrained at the Edge |
 | Non-idempotent device command (retried command doubles the effect) | P5 Edge Operations are Idempotent |
 | Blocking synchronous call on a constrained device with no timeout | P3, P5 (blocks the node; retry unsafe) |
 | Fire-and-forget telemetry with no on-device buffer (lost on partition) | P10 Edge Observability Survives Partition |
 | Whole-system crash on one device's failure (no degradation contract) | P7 Partial Degradation is Engineered |
 | Device credential scope that includes fleet-wide topics (over-scoped) | P9, `domains/security/secrets` |
 | QoS 2 used on a class-0 device (no budget for the handshake) | P3 Resources are Constrained and Declared |
 | Per-device-drop alert (noise; aggregate degradation is the signal) | P7, `domains/observability/metrics` |
 | Telemetry payload that includes device credentials or PII | P9, `domains/observability/P6 No Secrets in Observability` |
@@ -0,0 +1,354 @@
 # Offline-First — Derived Rules
 > Derives from `domains/edge/first-principles.md`. Applies P2
 > (Offline is a First-Class State) primarily, with P5 (idempotent
 > queue-and-forward), P4 (bounded sync conflicts on reconnect), P7
 > (partial degradation), and P10 (local-first telemetry). Cross-links
 > `domains/concurrency/patterns` for the in-process bounded-buffer
 > analog and `domains/observability/logging` for local-first logging.
 ## What Offline-First Is (P2 Offline is a First-Class State)
 - Offline-first is the design discipline in which the system
  continues to operate when disconnected from the center. Partition
  is the norm, not the exception; reconciliation happens on
  reconnect. The offline state is engineered, not a degenerate mode
  the app falls into by accident.
 - The boundary is per D-061: edge owns the
  proximity/location/disconnection angle. An offline-first web app
  is an edge concern because its defining trait is partition-survival
  (P2), not generic performance. The local-first storage is the
  edge device's constrained-resource reality (P3).
 - Offline-first is the precondition for the bounded-conflict
  discipline of `P4 Sync Conflicts are Bounded, Not Infinite`:
  without offline operation there is nothing to reconcile; with it,
  the reconnect reconciliation is the correctness mechanism. See
  `domains/edge/sync.md` for the conflict-resolution strategies.
 ## Local-First Storage (P2, P3)
 - Local-first storage holds the working copy on the device:
  IndexedDB (browser), SQLite (mobile, embedded), or on-device file
  storage (desktop, IoT gateway). The local store is the authority
  while offline; the server is reconciled later, not consulted per
  read.
 - The local store is bounded by the device (P3 — Resources are
  Constrained and Declared). A local store that grows without bound
  is a defect: declare a budget (e.g., a 50 MB IndexedDB quota, a
  30-day rolling window), and evict outside the budget deterministically.
 - The local store is the offline state; without it the app is
  online-only and crashes on disconnect (P2 violation). The store is
  the reversibility mechanism (C5): every local write is reversible
  on reconcile.
 ```typescript
 // Local-first store sketch (IndexedDB). The app reads from the
 // local store, never the network, while offline. Writes queue
 // locally and forward on reconnect (P2, P5).
 const db = await openDB("atelier-offline", 1, {
  upgrade(db) {
    const store = db.createObjectStore("pending-writes", {
      keyPath: "id",
    });
    store.createIndex("by-createdAt", "createdAt");
  },
 });
 async function readRecord(id: string) {
  // Read from local store first; the network is a reconcile path,
  // not the read path.
  return db.get("pending-writes", id);
 }
 ```
 ## Queue-and-Forward for Writes (P5 Edge Operations are Idempotent)
 - Every write while offline is queued locally and forwarded to the
  server on reconnect. The queue is the offline write-queue; the
  forward is the reconcile. Each queued write carries an idempotency
  key so a retried forward (the network is partition-prone) does not
  double-apply (P5).
 - The queue is bounded (P3): a queue that grows without limit on a
  constrained device will exhaust it. Declare a max-queue-depth and
  a max-queue-bytes; reject or evict beyond the bound with a defined
  policy (oldest-first, lowest-priority-first).
 - The queue is the cross-partition analog of the in-process bounded
  buffer — see `domains/concurrency/patterns` (bounded buffer,
  backpressure). Concurrency owns the in-process analog; edge owns
  the partition-survivable analog. The failure model differs: the
  in-process buffer fails by OOM; the offline write-queue fails by
  partition or device loss.
 ```typescript
 // Offline write-queue sketch. Each entry carries an idempotency
 // key (P5) so a retried forward is safe. The queue is bounded by
 // maxDepth (P3).
 interface PendingWrite {
  id: string;            // local id
  idempotencyKey: string; // server-side dedup key (P5)
  collection: string;
  payload: unknown;
  createdAt: number;
 }
 const MAX_DEPTH = 1000;
 async function queueWrite(write: Omit<PendingWrite, "id" | "idempotencyKey" | "createdAt">) {
  const depth = await db.count("pending-writes");
  if (depth >= MAX_DEPTH) {
    // P3: bounded queue. Evict the oldest pending write or reject.
    // Rejecting is correct when the write is higher-priority than
    // the oldest; evicting is correct when the newest is lowest.
    throw new Error("offline-queue-full");
  }
  const entry: PendingWrite = {
    ...write,
    id: crypto.randomUUID(),
    idempotencyKey: `${write.collection}:${crypto.randomUUID()}`,
    createdAt: Date.now(),
  };
  await db.put("pending-writes", entry);
  // The forward loop picks this up when connectivity returns.
 }
 async function forwardPendingWrites(server: Server) {
  const pending = await db.getAllFromIndex("pending-writes", "by-createdAt");
  for (const write of pending) {
    // P5: idempotent — the server dedups by idempotencyKey.
    await server.apply(write, write.idempotencyKey);
    await db.delete("pending-writes", write.id);
  }
 }
 ```
 ## Conflict Detection on Reconnect (P4 Sync Conflicts are Bounded)
 - On reconnect, the queued writes are forwarded; the server may
  have advanced while the device was offline. A conflict is when the
  local write and the server state diverge. Conflict detection is
  the precondition for bounded reconciliation (P4): a write forwarded
  blindly (last-write-wins with no clock) is a P4 violation waiting
  to happen.
 - Conflict resolution strategies (CRDT, LWW with vector clocks,
  application-specific merge) are the subject of
  `domains/edge/sync.md` — the CRDT-vs-LWW decision matrix there
  determines which applies. Offline-first owns the *detection*; sync
  owns the *resolution*.
 - A reconnect that detects no conflicts when conflicts exist is a
  silent correctness defect (C1, P4). Detection must be conservative:
  when in doubt, flag a conflict and surface it to the merge
  function or the user.
 ## UI for Offline State (P7 Partial Degradation is Engineered)
 - The UI must reflect the offline state visibly: a "you are offline,
  changes will sync when connected" banner, a pending-writes counter,
  a last-synced timestamp. A UI that hides the offline state
  violates P7 — the degraded mode is a designed state with a defined
  contract, not a silent fall-through.
 - The UI must function while offline: reads from local-first
  storage, writes to the queue, navigation that does not require the
  network. An app that shows a blank screen or a spinner-forever when
  offline has no offline state (P2 violation) and no degradation
  contract (P7 violation).
 - The pending-writes counter is the local-first analog of the
  messaging consumer-lag metric — see
  `domains/observability/metrics` for the lag-discipline parallel.
 ## Service Workers (P2, P6)
 - A service worker is a client-side proxy that intercepts network
  requests and serves from a local cache. It is the browser's
  offline-first primitive: the service worker cache is the
  offline-capable store for assets; the IndexedDB store is the
  offline-capable store for data.
 - The service worker cache is an edge cache (P6 — Cache Invalidation
  is Explicit): it must carry a TTL or explicit invalidation
  strategy. A service worker that caches forever and never
  invalidates is a TTL-less edge cache under partition — a P6
  violation (stale-forever).
 - See `domains/edge/cdn.md` for the generic edge-cache invalidation
  discipline; the service worker is the on-device instance of it.
 ```javascript
 // Service worker cache strategy: stale-while-revalidate for
 // assets, network-first for data, explicit version-bump for
 // breaking changes (P6).
 const CACHE = "atelier-v3";      // bump on deploy to invalidate (P6)
 const ASSETS = ["/", "/app.js", "/styles.css"];
 self.addEventListener("install", (event) => {
  event.waitUntil(
    caches.open(CACHE).then((cache) => cache.addAll(ASSETS))
  );
 });
 self.addEventListener("fetch", (event) => {
  const url = new URL(event.request.url);
  if (url.pathname.startsWith("/api/")) {
    // Network-first for data; fall back to cache on partition (P2).
    event.respondWith(
      fetch(event.request).catch(() => caches.match(event.request))
    );
  } else {
    // Stale-while-revalidate for assets (P6 explicit invalidation).
    event.respondWith(
      caches.open(CACHE).then(async (cache) => {
        const cached = await cache.match(event.request);
        const network = fetch(event.request).then((resp) => {
          cache.put(event.request, resp.clone());
          return resp;
        }).catch(() => cached);
        return cached || network;
      })
    );
  }
 });
 self.addEventListener("activate", (event) => {
  // P6: explicit invalidation. Drop old caches on activate.
  event.waitUntil(
    caches.keys().then((keys) =>
      Promise.all(keys.filter((k) => k !== CACHE).map((k) => caches.delete(k)))
    )
  );
 });
 ```
 ## Offline Write-Queue and Conflict Detection Mapped to the Testing Pyramid (IDEATE-38, ATELIER-94)
 The offline write-queue and conflict-detection patterns must be
 tested at every tier of the testing pyramid. Each tier exercises a
 different failure mode; skipping a tier leaves a correctness gap
 (P2, P4 violations that surface only in production partitions).
 | Pyramid Tier | What it exercises | What it proves |
 |-------------|-------------------|----------------|
 | **Unit** | Conflict detection on a merge function (pure inputs → expected merge result) | The merge logic is correct in isolation (P4) — given two divergent states, the merge returns the converged state |
 | **Integration** | Reconnect reconcile against a local store (fake server, real IndexedDB/SQLite) | The queue-and-forward loop drains correctly; the local store and server converge after reconnect (P2, P5) |
 | **E2e** | Partition simulation with a fake network (the app runs in a browser, the network is cut and restored) | The offline state, UI, and reconcile work end-to-end under partition (P2, P7) |
 - **Unit — conflict detection on a merge function.** The merge
  function is pure: given two divergent states and a clock, it
  returns the converged state. Test every merge case (LWW, CRDT
  register, set union, application-specific three-way merge) as a
  pure function. This is the cheapest tier and the highest coverage
  per test — see `domains/testing/pyramid`.
 ```typescript
 // Unit test sketch: conflict detection on a merge function (P4).
 // The merge function is pure; no network, no store. Test that
 // divergent states converge and that the merge is bounded (no
 // oscillation).
 function mergeLWW(local: State, remote: State, clock: Clock): State {
  // Last-write-wins: the state with the later vector-clock wins.
  // Returns the converged state (P4).
  return clock.compare(local.clock, remote.clock) >= 0 ? local : remote;
 }
 // Unit cases:
 // - local ahead → local wins
 // - remote ahead → remote wins
 // - concurrent (clocks incomparable) → conflict flagged or LWW tiebreak
 // - identical → no-op convergence (bounded, no oscillation)
 test("mergeLWW converges when local is ahead", () => {
  const local = { v: 2, clock: { a: 2 } };
  const remote = { v: 1, clock: { a: 1 } };
  expect(mergeLWW(local, remote, { compare: (a, b) => a.a - b.a })).toEqual(local);
 });
 ```
 - **Integration — reconnect reconcile against a local store.** A
  fake server stands in for the network; the real IndexedDB (or
  SQLite) holds the queue. The test fills the queue while offline,
  reconnects, and asserts the queue drains and the server and local
  store converge. This exercises the queue-and-forward loop (P5)
  and the reconcile against real storage.
 ```typescript
 // Integration test sketch: reconnect reconcile against a local
 // store. A fake server; real IndexedDB. The queue drains; the
 // server and local store converge after reconnect (P2, P5).
 test("reconnect reconciles pending writes against the server", async () => {
  const db = await openDB("test-offline", 1, { /* schema */ });
  const server = new FakeServer();
  await queueWrite(db, { collection: "docs", payload: { v: 1 } });
  // Simulate offline: server is unreachable.
  server.offline();
  await queueWrite(db, { collection: "docs", payload: { v: 2 } });
  expect(await db.count("pending-writes")).toBe(2);
  // Simulate reconnect: server is reachable.
  server.online();
  await forwardPendingWrites(server, db);
  expect(await db.count("pending-writes")).toBe(0);
  expect(await server.latest("docs")).toEqual({ v: 2 });
 });
 ```
 - **E2e — partition simulation with a fake network.** The app runs
  in a real browser; a fake network layer cuts and restores the
  connection. The test asserts the UI shows the offline state, the
  writes queue, the reconnect reconciles, and the UI returns to
  online. This is the highest-fidelity tier and the lowest coverage
  per test — run a small number of representative scenarios, not a
  combinatorial matrix.
 ```typescript
 // E2e test sketch: partition simulation with a fake network. The
 // app runs in a browser; the network is cut and restored. Asserts
 // the offline UI state, the queue, the reconcile, and the online
 // recovery (P2, P7).
 test("app survives a network partition and reconciles on reconnect", async () => {
  await page.goto("https://app.example.com");
  await page.click("text=Edit document");
  await page.fill("textarea", "offline edit");
  // Cut the network.
  await page.setOffline(true);
  await page.click("text=Save");
  await expect(page.locator("text=You are offline")).toBeVisible();
  await expect(page.locator("text=1 pending change")).toBeVisible();
  // Restore the network.
  await page.setOffline(false);
  await expect(page.locator("text=All changes synced")).toBeVisible();
  await expect(page.locator("text=0 pending changes")).toBeVisible();
 });
 ```
 - The three tiers are complementary: unit proves the merge logic,
  integration proves the reconcile loop, e2e proves the partition
  behavior. Skipping any tier leaves a correctness gap. See
  `domains/testing/pyramid` for the pyramid discipline and
  `domains/testing/fixtures` for the fake-server and fake-network
  fixture patterns.
 ## Observability (P10 Edge Observability Survives Partition)
 - The offline state is itself an observable signal: the
  pending-writes count, the last-synced timestamp, the
  reconcile-failure count. A device stuck offline for days with a
  full queue is an incident; without local-first telemetry it is
  invisible (P10 violation).
 - Local-first logging (buffered on-device, forwarded on reconnect)
  is the offline-first instance of `P10 Edge Observability Survives
  Partition`. See `domains/observability/logging` for the generic
  structured-logging discipline the local-first buffer builds on.
 - A reconcile failure that is not logged locally is a silent defect
  — the operator cannot debug what they cannot see (C7, P10).
 ## What Violates Offline-First Discipline
 | Violation | Principle |
 |-----------|-----------|
 | App that crashes on disconnect (no offline state) | P2 Offline is a First-Class State |
 | Unbounded offline write-queue (grows until device OOM) | P3 Resources are Constrained and Declared |
 | Queued write forwarded without an idempotency key (retry doubles the effect) | P5 Edge Operations are Idempotent |
 | Reconnect that detects no conflicts when conflicts exist | P4 Sync Conflicts are Bounded, Not Infinite |
 | UI that hides the offline state (no banner, no pending counter) | P7 Partial Degradation is Engineered |
 | Service worker cache with no TTL and no explicit invalidation | P6 Cache Invalidation is Explicit |
 | Reconcile failure with no local log (silent under partition) | P10 Edge Observability Survives Partition |
 | Merge function that oscillates (no convergence guarantee) | P4, `domains/edge/sync.md` |
 | Local-first store with no declared budget (grows without bound) | P3, `domains/concurrency/patterns` (bounded buffer analog) |
 | E2e tests that never simulate a partition (offline path untested) | P2, `domains/testing/pyramid` |
@@ -0,0 +1,315 @@
 # Sync — Derived Rules
 > Derives from `domains/edge/first-principles.md`. Applies P4 (Sync
 > Conflicts are Bounded, Not Infinite) primarily, with P5
 > (idempotent merge operations), P2 (offline as the precondition),
 > and P10 (sync is observable). Cross-links `domains/data/migrations`
 > for schema migration under sync and `domains/concurrency/patterns`
 > for the immutability-aid-merge principle.
 ## What the Sync Problem Is (P4 Sync Conflicts are Bounded, Not Infinite)
 - Sync is the discipline of reconciling divergent state across
  partitioned nodes. While partitioned, each node accepts writes
  independently; on reconnect, the divergent state must converge.
  The correctness contract is that the merge terminates and
  converges — oscillation and infinite sync loops are correctness
  failures, not eventual consistency (P4).
 - Sync is the edge domain's deepest problem: it is the
  reconciliation layer above `P2 Offline is a First-Class State`.
  Without offline operation there is nothing to sync; with it, the
  reconnect reconciliation is the correctness mechanism. See
  `domains/edge/offline-first.md` for the offline write-queue that
  produces the divergent state to be reconciled.
 - The boundary is per D-061: edge owns the partitioned-reconcile
  angle; concurrency owns the in-process analog
  (`concurrency/P1 Immutability by Default` — immutability aids
  merge); data owns the generic migration discipline
  (`data/migrations`). Sync is an edge concern because its defining
  trait is partitioned divergence, a concern that only arises at the
  network edge.
 ## Conflict Resolution Strategies (P4, C1 Correctness, C5 Reversibility)
 - A conflict is when two nodes have divergent state for the same
  logical entity and no total order determines which is correct.
  Resolution strategies fall into two families:
  - **Conflict-free**: the data type guarantees convergence by
    construction (CRDTs). The merge is deterministic; no conflict
    surfaces to the user or the application.
  - **Conflict-tolerant**: the data type can conflict; the
    resolution policy (last-write-win, three-way merge,
    application-specific) arbitrates. Conflicts may surface to the
    user or be silently resolved per a documented policy.
 - The choice is a P4 decision: conflict-free types guarantee the
  bound (convergence) but constrain the data model; conflict-tolerant
  types are flexible but require the resolution policy to be correct
  and bounded (no oscillation). See the decision matrix below.
 ## CRDTs — Conflict-Free Replicated Data Types (P4, C5, C6)
 - A CRDT is a data type whose merge operation is associative,
    commutative, and idempotent. Given any set of divergent replicas,
  merging them in any order converges to the same state — the merge
  is deterministic and terminating (P4 bound). CRDTs derive from C5
  Reversibility (divergent state reverses to convergence) and C6
  Composability (CRDTs compose: a CRDT map of CRDT registers is
  itself a CRDT).
 - **State-based (CvRDT — convergent):** each replica carries its
  full state; merge is a least-upper-bound on a semi-lattice. The
  payload is larger (full state per merge); the merge is simple
  (one function). Fits small state and unreliable networks.
 - **Operation-based (CmRDT — commutative):** each replica carries
  operations; merge is applying the operations in causal order. The
  payload is smaller (ops, not state); the delivery must be
  reliable and causally ordered. Fits large state and reliable
  transport.
 - The tradeoff: state-based is simpler but heavier; operation-based
  is lighter but requires causal delivery. Both guarantee
  convergence (P4); the choice is a C8 Economy decision (bandwidth
  vs delivery complexity).
 ```typescript
 // CRDT register: LWW-element-set (state-based, CvRDT). The merge
 // is deterministic — the register with the later timestamp wins.
 // Convergence is guaranteed (P4); the merge is idempotent (P5).
 interface LWWRegister<T> {
  value: T;
  timestamp: number;   // monotonic clock; ties broken by node id
  nodeId: string;
 }
 function mergeLWWRegister<T>(
  local: LWWRegister<T>,
  remote: LWWRegister<T>,
 ): LWWRegister<T> {
  // The merge is associative, commutative, idempotent (P4, P5).
  // (local.timestamp, local.nodeId) > (remote.timestamp, remote.nodeId)
  // is a total order — no oscillation, no infinite loop.
  if (local.timestamp > remote.timestamp) return local;
  if (local.timestamp < remote.timestamp) return remote;
  // Tie: break by node id for a deterministic total order.
  return local.nodeId > remote.nodeId ? local : remote;
 }
 // The register is a CRDT: merge(merge(a, b), c) === merge(a, merge(b, c))
 // for any replicas a, b, c. Convergence is guaranteed (P4).
 ```
 ```typescript
 // CRDT set: add-wins last-write-wins element set (state-based).
 // Each element carries a timestamp; remove only wins if the
 // remove-timestamp is later than the add-timestamp. This avoids
 // the remove-wins-vs-add race (P4) without surfacing a conflict.
 interface AWLWWSet<T> {
  adds: Map<T, number>;     // element -> add-timestamp
  removes: Map<T, number>;  // element -> remove-timestamp
 }
 function mergeAWLWWSet<T>(a: AWLWWSet<T>, b: AWLWWSet<T>): AWLWWSet<T> {
  const adds = new Map<T, number>(a.adds);
  const removes = new Map<T, number>(a.removes);
  for (const [el, ts] of b.adds) {
    adds.set(el, Math.max(adds.get(el) ?? 0, ts));      // add-wins union
  }
  for (const [el, ts] of b.removes) {
    removes.set(el, Math.max(removes.get(el) ?? 0, ts));
  }
  return { adds, removes };
 }
 function contains<T>(set: AWLWWSet<T>, el: T): boolean {
  const addTs = set.adds.get(el) ?? 0;
  const rmTs = set.removes.get(el) ?? 0;
  return addTs > rmTs;   // add wins on equal timestamp (P4 bounded)
 }
 ```
 ## Last-Write-Win (LWW) with Vector Clocks (P4, C1, C5)
 - LWW is the simplest conflict-tolerant strategy: the write with the
  latest timestamp wins. It is cheap, but it silently discards
  concurrent writes — the "lost update" is the correctness cost. LWW
  is correct only when the timestamp is a total order (a monotonic
  clock, not wall time), and when lost concurrent writes are
  acceptable (e.g., caching, presence, ephemeral state).
 - **Vector clocks** are the timestamp that knows about concurrency.
  A vector clock records the logical time of each node; two writes
  are concurrent iff neither vector dominates the other. LWW with
  vector clocks: a write that is causally later wins; a write that
  is concurrent conflicts and is resolved by a tiebreak (node id,
  wall time, or application policy).
 - The tiebreak is the P4 bound: the conflict must be resolved
  deterministically (no oscillation) and the resolution must be
  documented. A tiebreak by wall time alone (no vector clock) is a
  P4 violation waiting to happen — wall time skews across nodes,
  and a clock skew can flip the tiebreak, oscillating the merge.
 ```typescript
 // LWW with vector clocks (conflict-tolerant, P4 bounded). The
 // vector clock records causal order; concurrent writes conflict;
 // the conflict is tiebroken deterministically (no oscillation).
 type VectorClock = Record<string, number>;  // nodeId -> counter
 function compareClock(a: VectorClock, b: VectorClock): "before" | "after" | "equal" | "concurrent" {
  let aBefore = false, bBefore = false;
  const keys = new Set([...Object.keys(a), ...Object.keys(b)]);
  for (const k of keys) {
    const av = a[k] ?? 0;
    const bv = b[k] ?? 0;
    if (av < bv) aBefore = true;
    if (av > bv) bBefore = true;
  }
  if (aBefore && bBefore) return "concurrent";
  if (aBefore) return "before";
  if (bBefore) return "after";
  return "equal";
 }
 interface LWWVectorState<T> {
  value: T;
  clock: VectorClock;
  writerId: string;  // tiebreak: deterministic, no oscillation (P4)
 }
 function mergeLWWVector<T>(
  local: LWWVectorState<T>,
  remote: LWWVectorState<T>,
 ): LWWVectorState<T> {
  const order = compareClock(local.clock, remote.clock);
  if (order === "before") return remote;       // remote causally later
  if (order === "after" || order === "equal") return local;
  // Concurrent: tiebreak by writer id (deterministic, P4 bounded).
  return local.writerId > remote.writerId ? local : remote;
 }
 ```
 - The merge is idempotent (P5): merging the same two replicas twice
  yields the same result. The tiebreak by `writerId` is a total
  order, so the merge cannot oscillate (P4 bound).
 - A vector-clock merge that surfaces the concurrent conflict to the
  application (instead of tiebreaking) is also valid — the
  application resolves per its own policy. The P4 bound is that the
  resolution terminates; the policy determines whether the user sees
  the conflict or the system silences it.
 ## Merge Semantics (P4, P5, cross-link concurrency/patterns)
 - The merge function is the heart of sync. Its properties (P4):
  - **Associative**: `merge(merge(a, b), c) === merge(a, merge(b, c))`.
  - **Commutative**: `merge(a, b) === merge(b, a)`.
  - **Idempotent**: `merge(a, a) === a` (P5 — retried merges are safe).
 - Immutability aids merge: an immutable state representation (the
  CRDT payload, the LWW register with a clock) makes the merge a
  pure function of two inputs, with no in-place mutation race. See
  `domains/concurrency/patterns` (`concurrency/P1 Immutability by
  Default`) for the in-process immutability principle; sync is the
  cross-partition instance of it.
 - A merge that mutates in place is a P5 violation waiting to
  happen: a retried merge mutates the same state twice, and the
  result is not idempotent. Always merge into a new state; never
  mutate the inputs.
 ## Conflict-Free vs Conflict-Tolerant Data Types (P4, C3 Simplicity)
 - **Conflict-free (CRDTs):** the data type guarantees convergence.
  The application never sees a conflict; the merge is deterministic.
  The cost: the data model is constrained (counters, sets, registers,
  maps of these). A conflict-free type for arbitrary JSON is hard;
  a conflict-free type for a counter is a PN-counter.
 - **Conflict-tolerant (LWW, three-way merge, application policy):**
  the data type can conflict; the resolution policy arbitrates. The
  cost: the policy must be correct and bounded (no oscillation), and
  the conflict may surface to the user. The benefit: any data model
  can be made conflict-tolerant (just pick a tiebreak).
 - The choice is the decision matrix below. It is a P4 decision
  (which bound), a C1 decision (which correctness cost is
  acceptable), and a C3 decision (which simplicity is affordable).
  See also `domains/data/migrations` for the schema-evolution angle
  — a schema change under sync must be compatible with both
  replicas, or the merge fails on the new shape.
 ## Schema Migration Under Sync (P4, cross-link data/migrations)
 - A schema migration under sync is harder than a single-node
  migration: both replicas must understand the new shape, or the
  merge fails. The migration must be forward-and-backward compatible
  across all replicas that may still hold the old shape — see
  `domains/data/migrations` for the generic compatibility discipline.
 - A breaking schema change under sync requires a staged migration:
  deploy the new-shape-aware merge first (it accepts both shapes),
  then deploy the new shape, then deploy the old-shape-removing
  merge. A big-bang schema change under sync is a P4 violation: the
  replicas that have not yet upgraded will fail the merge, and the
  sync will not converge.
 - The merge function's version awareness is the P4 bound: the merge
  must handle every shape version that may exist in the fleet, or
  reject (and surface) the merge rather than silently corrupting.
 ## CRDT vs Last-Write-Win — Decision Matrix (D-069)
 | Strategy | When | Correctness Guarantee | Operational Cost | Failure Mode |
 |----------|------|------------------------|-------------------|--------------|
 | CRDT (state-based, CvRDT) | The data model fits a CRDT (counter, set, register, map of these); convergence must be guaranteed without surfacing conflicts; the network is unreliable (full-state merge tolerates dropped ops) | Strong eventual convergence — `merge(a, b) === merge(b, a)` for any replicas (P4 bound by construction) | Medium — full state per merge (bandwidth); semi-lattice merge function per type; CRDT library or hand-rolled | A bug in the merge function = silent divergence (C1); large state = bandwidth cost on constrained links (P3) |
 | CRDT (operation-based, CmRDT) | The data model fits a CRDT; bandwidth is constrained (ops are smaller than state); the transport is reliable and causally ordered | Strong eventual convergence — same guarantee, smaller payload | High — requires causal delivery (vector clock or broker with ordering); op transform must be idempotent (P5) | Causal-delivery violation = lost ops = divergence; op-transform bug = silent divergence |
 | Last-Write-Win (LWW) with vector clocks | The data model is arbitrary (any JSON, any record); concurrent writes are acceptable to discard or tiebreak; a total order tiebreak (node id) is acceptable | Bounded convergence — causally-later writes win; concurrent writes are tiebroken deterministically (P4 bound via tiebreak) | Low — simple merge (compare clocks, pick winner); no CRDT library; small payload | Concurrent writes are silently discarded (lost update); tiebreak by wall time = clock-skew oscillation (P4 violation); no vector clock = no concurrent-write detection = silent loss |
 | LWW with wall-clock timestamp only | The data model is ephemeral (cache, presence); lost updates are acceptable; the clock is roughly synchronized (NTP) | Weak — convergence eventually, but concurrent writes may oscillate with clock skew; no concurrent-write detection | Lowest — one timestamp per write; no clock vector | Clock skew = oscillation (P4 violation); concurrent writes silently lost; not a correctness-safe strategy for durable state |
 | Three-way merge (application-specific) | The data model is structured (documents, forms); conflicts should surface to the user or a domain-specific resolver; the merge is field-level | Bounded if the merge function is correct (associative, commutative, idempotent — P4, P5); conflicts surface per field | High — application-specific merge function per type; UI for conflict resolution; user-facing conflict surface | Merge-function bug = silent divergence or oscillation; unbounded conflict UI = user fatigue |
 - The default for structured state that must converge silently is a
  **CRDT** (state-based for unreliable networks, operation-based for
  bandwidth-constrained reliable transport). The default for
  arbitrary JSON where lost concurrent updates are acceptable is
  **LWW with vector clocks** (never wall-clock-only for durable
  state). The default for user-facing documents where conflicts
  should surface is **three-way merge** with a documented resolution
  policy.
 - The failure-mode column is the P4 check: every row except
  wall-clock-only LWW carries a bounded failure mode (the bug is in
  the implementation, not the strategy). Wall-clock-only LWW carries
  an unbounded failure mode (clock skew = oscillation) and is a P4
  violation for durable state. Use it only for ephemeral state
  where lost updates are acceptable.
 - The choice is a P4 decision (which bound) and a C1 decision
  (which correctness cost). A CRDT guarantees convergence but
  constrains the data model; LWW is flexible but discards concurrent
  writes. Neither is universally correct; the matrix is the
  decision tool.
 ## Observability of Sync (P10 Edge Observability Survives Partition)
 - Sync is itself an observable operation: the merge count, the
  conflict count, the convergence lag (time from reconnect to
  convergence), and the divergent-replica count are first-class
  signals. A sync that runs forever without converging is the
  `edge-sync-loop` chaos anti-pattern (P4 breach); without
  observability it is invisible until the user notices the stale
  state.
 - A conflict that is silently resolved should be logged (the
  resolution policy applied, the discarded write's idempotency key,
  the winning write's clock). A conflict that surfaces to the user
  should be metricated (the conflict rate, the resolution time).
  See `domains/observability/metrics` for the generic discipline.
 - A divergent replica that has not converged after the expected
  window is an incident; without a metric it is invisible (P10
  breach). Wire sync convergence to an alert — the
  `divergent-replica-count` is the sync analog of the messaging
  `consumer-lag` metric.
 ## What Violates Sync Discipline
 | Violation | Principle |
 |-----------|-----------|
 | Sync loop that oscillates forever (CRDT without merge-semantics, LWW without monotonic clock) | P4 Sync Conflicts are Bounded, Not Infinite |
 | LWW with wall-clock timestamp only on durable state (clock skew = oscillation) | P4, C1 (no concurrent-write detection) |
 | Merge function that mutates inputs in place (retried merge is not idempotent) | P5 Edge Operations are Idempotent, `domains/concurrency/patterns` |
 | Big-bang schema change under sync (replicas fail the merge) | P4, `domains/data/migrations` |
 | Conflict silently resolved with no log (the policy is invisible) | P10 Edge Observability Survives Partition |
 | Divergent replica with no convergence-lag metric (invisible stale state) | P10, `domains/observability/metrics` |
 | Three-way merge with an unbounded conflict UI (user fatigue, no termination) | P4 (the merge must terminate) |
 | Operation-based CRDT without causal delivery (lost ops = divergence) | P4, C1 (the delivery contract is the bound) |
 | Merge that surfaces every concurrent conflict to the user (no default policy) | P4, C3 (the default policy is the simplicity bound) |
 | Sync with no convergence test (the merge is untested under partition) | P4, `domains/testing/pyramid` |
@@ -0,0 +1,173 @@
 # Internationalization (i18n) — First Principles
 > Grounded in Unicode ICU + CLDR, W3C i18n WG, BCP 47 / RFC 5646,
 > ICU MessageFormat / FormatJS / i18next / Mozilla Fluent, the
 > JavaScript `Intl` API, and WCAG 2.1 AA. The developer's language is
 > one locale among many, not the neutral form.
 ## 1. The Principles
 ### P1. Source Language is a Locale, Not the Default
 The developer's own language is one locale among many — it is not the
 "neutral" or "unlocalized" form of the product. Strings are extracted
 from day one, addressed by key, and routed through a locale resource
 layer even when only one locale is populated. Treating the source
 language as the default produces hidden concatenations, hardcoded
 grammar assumptions, and a translation debt that compounds until the
 first second locale arrives — at which point the fix is a rewrite, not
 a patch. The source locale is `en-US` (or whatever the team writes in);
 it is not `null`. This is the i18n angle on `domains/uiux/copywriting.md`:
 copy lives in resources, not in code.
 ### P2. Locale Identifiers are Standardized
 Use BCP 47 language tags (`en-US`, `ar-EG`, `zh-Hans-CN`, `pt-BR`).
 No ad-hoc locale codes, no two-letter-only hacks, no invented keys.
 The tag carries language, script (when needed), and region (when
 needed); it is the contract between the resource layer, the
 formatting layer, and the runtime. A locale identifier that is not
 BCP 47 is a key that cannot be resolved by any standard tool, which
 is a correctness violation. Cross `domains/data/schema-design.md`:
 locale identifiers are a data shape with a defined vocabulary.
 ### P3. Resources are External, Not Inline
 User-facing strings live in locale resource files (`.po`, JSON,
 Fluent `.ftl`, ICU Resource Bundle), never concatenated inline in
 code. Inline strings are invisible to the translation pipeline,
 unversionable as a unit, and untestable for completeness. String
 concatenation in code (`"Welcome, " + name + "!"`) is the cardinal
 violation: it bakes in source-language grammar and breaks for every
 locale with different word order. Resources are the boundary; code
 addresses strings by key, the resource layer resolves the key to the
 locale. This is the i18n angle on C4 Locality: strings and their
 locale-specific consequences live together in the resource, not
 scattered across code.
 ### P4. Plural and Gender are Parameterized
 Plural forms, gender, and select are expressed with ICU MessageFormat
 (or an equivalent parameterized formatter), never with `if (n == 1)`
 branching in code. Plural rules are locale-specific — English has
 one/other, Arabic has six categories, Russian has three — and a
 hand-rolled branch encodes exactly one locale's rules while pretending
 to be universal. The formatter is the contract; the resource carries
 the variants; the code passes the count and lets the formatter choose.
 A `if (n == 1)` plural is a C1 (Correctness) violation masquerading as
 a shortcut.
 ### P5. Formatting is Locale-Aware
 Dates, times, numbers, currencies, units, and relative time are
 formatted via ICU / CLDR / the JavaScript `Intl` API — never
 hand-rolled. A hand-rolled date formatter encodes one locale's
 conventions and silently produces wrong output for every other locale
 (mm/dd/yyyy vs dd/mm/yyyy is the canonical failure). CLDR is the
 source of truth for locale data; `Intl` is the runtime that exposes
 it. Formatting correctness is observable: a misformatted date is a
 wrong answer in the user's locale, even if it is "right" in the
 developer's. Cross `domains/api/error-responses.md` for localized
 error messages at API boundaries.
 ### P6. Text Direction is a Layout Primitive
 RTL and bidi are first-class layout concerns, not a CSS afterthought.
 Logical CSS properties (`margin-inline-start`, `padding-block-end`,
 `inset-inline-end`) over physical (`margin-left`, `padding-top`). The
 `dir` attribute is set on the document and on subtrees; the bidi
 algorithm (UAX #9) handles inline reordering. A layout that assumes
 LTR is a layout that is wrong for `ar`, `he`, `fa`, `ur`, and any
 RTL-mixed context. Text direction is not a skin — it is a structural
 property of the layout, and fixing it late is a rewrite. This is the
 i18n angle on `domains/uiux/accessibility.md`: RTL support is an
 accessibility concern for non-Latin-script users.
 ### P7. Layout Accommodates Expansion
 Translated text expands and contracts — German is ~30% longer than
 English, Japanese often shorter, RTL mirroring shifts every visual
 anchor. Layouts are flexible: no fixed pixel widths for text, no
 truncation without an ellipsis-and-title strategy, no
 `white-space: nowrap` on translatable strings. A layout that breaks
 on a 30% expansion is a layout that is wrong for most of the world's
 locales. Designing for the worst-case expansion up front is cheaper
 than reworking every screen when the first long-form locale ships.
 ### P8. Pseudo-Locales Test Early
 Test with pseudo-locales (accented, lengthened, RTL-mirrored, brack-
 enclosed) before real translations arrive. A pseudo-locale run
 surfaces hardcoded strings, layout overflow, broken concatenation,
 and LTR assumptions while the fix is still cheap — the translator
 hasn't been paid yet, and the string freeze hasn't happened. Finding
 these bugs after real translation is a C5 (Reversibility) violation:
 the cost of undoing is now a re-translation. Cross
 `domains/testing/fixtures.md` and `domains/testing/pyramid.md` for
 where pseudo-locales sit in the testing pyramid.
 ### P9. Images and Icons are Cultural
 Icons, colors, gestures, and imagery are locale-sensitive. A
 mailbox icon means "email" in the US and "mail" in Japan — but a
 green checkmark means "correct" in the West and "incorrect" in some
 East Asian contexts. A thumbs-up is positive in much of the world
 and an insult in parts of the Middle East. Avoid locale-bound symbols
 as universal; parameterize imagery per locale where the symbol is not
 globally neutral. Icons are not a universal language; they are a
 locale with a picture. This is a C2 (Clarity) concern: an icon whose
 meaning changes by locale is unclear to the reader it was not drawn
 for.
 ### P10. Translation is Reversible and Versioned
 Resource files are versioned alongside code; a bad translation is a
 rollback, not a hot-patch. Every locale resource has a history
 (what shipped when), a provenance (which translator / which service),
 and a rollback path. A translation that breaks the UI is reverted to
 the prior resource version, the same way a code regression is
 reverted to the prior commit. Translations without version history
 are anecdote, not artifact — you cannot tell what changed, when, or
 why. This is the i18n angle on C5 Reversibility applied to the
 resource layer.
 ## 2. Core Principle Trace
 Each i18n P-rule derives from one or more core C-rules (C1–C8). The
 matrix extension lands in P4 of the v0.3 plan; the traces below are
 authoritative.
 | P-rule | Core | Why |
 |--------|------|-----|
 | P1 Source Language is a Locale, Not the Default | C2, C1 | Clarity of locale intent; correctness of treating source as one-of-many |
 | P2 Locale Identifiers are Standardized | C2, C6 | Clarity of a standard vocabulary; composability with standard tools |
 | P3 Resources are External, Not Inline | C4, C6 | Locality of strings and their locale consequences; composability of the resource layer |
 | P4 Plural and Gender are Parameterized | C1, C6 | Correctness of locale-specific plural rules; composability of the formatter contract |
 | P5 Formatting is Locale-Aware | C1, C7 | Correctness of formatted output; observability of format correctness |
 | P6 Text Direction is a Layout Primitive | C1, C4 | Correctness of layout for RTL; locality of direction with the text it governs |
 | P7 Layout Accommodates Expansion | C8, C3 | Economy of rework; simplicity of flexible layouts over per-locale overrides |
 | P8 Pseudo-Locales Test Early | C7, C5 | Observability of i18n defects early; reversibility of fixing before translation |
 | P9 Images and Icons are Cultural | C1, C2 | Correctness of locale-appropriate symbols; clarity of meaning across locales |
 | P10 Translation is Reversible and Versioned | C5 | Reversibility of the resource layer |
 ## 3. What Violates These Principles
 | Violation | Principle Breached |
 |-----------|-------------------|
 | A user-facing string hardcoded in source | P3 Resources are External, Not Inline |
 | `"Welcome, " + name + "!"` string concatenation | P3 Resources are External, Not Inline |
 | `if (n == 1) { return "item"; } else { return "items"; }` | P4 Plural and Gender are Parameterized |
 | A locale code like `en_us` or `english` instead of `en-US` | P2 Locale Identifiers are Standardized |
 | A hand-rolled date formatter (`getMonth() + 1 + "/" + getDay()`) | P5 Formatting is Locale-Aware |
 | `margin-left: 10px` on a translatable layout | P6 Text Direction is a Layout Primitive |
 | A fixed-width text container that overflows on German | P7 Layout Accommodates Expansion |
 | First i18n test runs against real translations, not pseudo-locales | P8 Pseudo-Locales Test Early |
 | A thumbs-up icon shipped as universally positive | P9 Images and Icons are Cultural |
 | Resource files with no git history or no rollback path | P10 Translation is Reversible and Versioned |
 | The source language treated as the "unlocalized" default | P1 Source Language is a Locale, Not the Default |
 ## 4. Relationship to Other Domains
 i18n is the locale-awareness layer that `domains/uiux/` consumes and
 that `domains/api/` surfaces at boundaries. It borrows the testing
 discipline of `domains/testing/` and the data-shape discipline of
 `domains/data/`. Cross-links are one-directional (per D-026 extended):
 - `domains/uiux/copywriting.md` ← P1, P3 (strings live in resources)
 - `domains/uiux/accessibility.md` ← P6 (RTL is an a11y concern for non-Latin users)
 - `domains/uiux/components.md` ← P6, P7 (layout primitives that survive direction and expansion)
 - `domains/api/error-responses.md` ← P5 (localized error messages)
 - `domains/data/schema-design.md` ← P2, P3 (locale data shapes)
 - `domains/testing/fixtures.md` ← P8 (pseudo-locale fixtures)
 - `domains/testing/pyramid.md` ← P8 (pseudo-locale tier mapping)
 - `domains/testing/first-principles.md` ← P8 (testing discipline for locale)
@@ -0,0 +1,141 @@
 # Formatting — Derived Rules
 > Derives from `domains/i18n/first-principles.md`. Covers P2 (Locale
 > Identifiers Standardized), P4 (Plural/Gender Parameterized), and P5
 > (Formatting is Locale-Aware). Referenced by `locale-resources.md`
 > (the formatter resolves the messages) and `testing-i18n.md` (the
 > formatted output is what snapshots assert).
 ## Formatting is Locale-Aware (P5 Formatting is Locale-Aware)
 - Dates, times, numbers, currencies, units, and relative time are
  formatted via ICU / CLDR / the JavaScript `Intl` API — never
  hand-rolled. CLDR is the source of truth for locale data; `Intl`
  is the runtime that exposes it.
 - A hand-rolled formatter encodes one locale's conventions and
  silently produces wrong output for every other locale. The
  canonical failure is date format: `mm/dd/yyyy` (US) vs
  `dd/mm/yyyy` (most of the world) vs `yyyy-mm-dd` (ISO, sortable).
  Picking one and calling it done is a correctness violation in
  every locale it is wrong for.
 ## BCP 47 Tags Drive Formatting (P2 Locale Identifiers Standardized)
 - Every formatter takes a BCP 47 locale tag. The tag is the contract
  between the resource layer and the formatting layer: the same tag
  that selects the resource selects the formatter.
 - A locale tag that is not BCP 47 cannot be resolved by `Intl`, ICU,
  or CLDR — the formatter returns the runtime default, which is the
  developer's locale, not the user's. This is why P2 is a
  prerequisite of P5: you cannot format for a locale you cannot name.
 ## The Intl Surface (ICU/CLDR in the Browser and Node)
 | API | Formats | Notes |
 |-----|---------|-------|
 | `Intl.DateTimeFormat` | Dates, times, date+time, time zones | Calendar (`buddhist`, `hebrew`, `islamic`), numbering system (`arab`, `hanidec`) via locale tag extensions |
 | `Intl.NumberFormat` | Numbers, currencies, units, percent | Notation (`compact`, `scientific`), grouping, sign display |
 | `Intl.RelativeTimeFormat` | "3 days ago", "in 2 months" | Locale-specific phrasing; numeric vs auto |
 | `Intl.PluralRules` | Plural category for a count | `one`, `few`, `many`, `other`, `zero`, `two` per CLDR — the engine ICU MessageFormat uses |
 | `Intl.ListFormat` | "a, b, and c" | Conjunction / disjunction / unit lists, locale-specific separators |
 | `Intl.Collator` | Locale-aware string sorting | Strength (`base`, `accent`, `case`); numeric collation |
 - All of these are built on ICU/CLDR; they are the runtime baseline.
  Use them. A `moment.js`-style hand-rolled format string
  (`"MM/DD/YYYY"`) is a relic of the pre-`Intl` era and a P5
  violation in any locale-aware code path.
 ## Dates and Times
 ```
 // Correct — Intl, locale-aware
 new Intl.DateTimeFormat("ar-EG", {
  dateStyle: "full",
  timeStyle: "short",
 }).format(new Date());
 // "الأربعاء، ٧ نوفمبر ٢٠٢٤، ٣:١٥ م"
 // Wrong — hand-rolled, source-locale only
 const d = new Date();
 const s = (d.getMonth() + 1) + "/" + d.getDate() + "/" + d.getFullYear();
 // "11/7/2024" — meaningless in most locales
 ```
 - Time zones are not locales. A locale tells you *how to format* a
  timestamp; a time zone tells you *what instant* it refers to. Do
  not derive one from the other (`ar-EG` is not a time zone).
  Format with the user's locale; render in the user's time zone;
  store in UTC.
 ## Numbers, Currencies, Units
 ```
 new Intl.NumberFormat("de-DE", { style: "currency", currency: "EUR" })
  .format(1234.56); // "1.234,56 €"
 new Intl.NumberFormat("ar-EG", { style: "currency", currency: "EGP" })
  .format(1234.56); // "١٬٢٣٤٫٥٦ ج.م.‏"
 new Intl.NumberFormat("en-US", { style: "unit", unit: "kilometer-per-hour" })
  .format(100); // "100 km/h"
 ```
 - The currency code (`EUR`, `EGP`, `USD`) is ISO 4217; the locale
  determines the symbol, grouping, and placement. A hand-rolled
  `"$" + amount` is wrong for `de-DE` (symbol, grouping, placement
  all differ).
 ## Plural Rules (P4 Plural/Gender Parameterized)
 - `Intl.PluralRules` returns the CLDR plural category for a count in
  a given locale. ICU MessageFormat uses this category to select the
  variant from the resource (`locale-resources.md`).
 - Never branch on the raw count in code. The count goes to the
  formatter; the formatter consults `PluralRules` for the locale;
  the resource carries the variant for that category.
  ```
  // ICU MessageFormat (FormatJS)
  new Intl.MessageFormat(
    "{count, plural, one {# item} other {# items}}",
    "en-US"
  ).format({ count: 1 }); // "1 item"
  // ar-EG — six categories; the code is identical, only the
  // resource differs.
  ```
 ## Gender and Select
 - ICU MessageFormat also supports `{gender, select, male {...} female {...} other {...}}`
  for gendered agreement and `{case, select, ...}` for general
  disjunction. These live in the resource, not in code branches.
 - A `switch (gender)` in code that picks a string is the same
  violation as `if (n == 1)`: it encodes one locale's grammar in
  code and breaks for every locale with different agreement rules.
 ## What Violates Formatting Discipline
 | Violation | Principle |
 |-----------|-----------|
 | `getMonth() + 1 + "/" + getDay()` hand-rolled date | P5 Formatting is Locale-Aware |
 | `"$" + amount` hand-rolled currency | P5 Formatting is Locale-Aware |
 | `if (n === 1) "item" else "items"` plural branch | P4 Plural and Gender are Parameterized |
 | `moment("MM/DD/YYYY")` format string in locale-aware code | P5 Formatting is Locale-Aware |
 | Deriving time zone from locale tag | P2 Locale Identifiers are Standardized |
 | A non-BCP-47 tag passed to `Intl` (silently falls back) | P2 Locale Identifiers are Standardized |
 | `switch (gender)` selecting strings in code | P4 Plural and Gender are Parameterized |
 | Storing timestamps in local time, not UTC | P5 Formatting is Locale-Aware |
 ## Relationship to Other Domains
 - `domains/api/error-responses.md` — API error messages are
  formatted for the requesting locale; the error code is stable, the
  message is locale-formatted.
 - `domains/data/schema-design.md` — locale identifiers, currency
  codes, and time zones are data contracts; treat them as schema
  (`en-US`, `EUR`, `UTC`), not free text.
 - `domains/i18n/locale-resources.md` — the resource layer carries
  the parameterized messages this formatter resolves.
 - `domains/testing/fixtures.md` — formatted output per locale is the
  fixture; snapshot tests assert against it.
@@ -0,0 +1,138 @@
 # Locale Resources — Derived Rules
 > Derives from `domains/i18n/first-principles.md`. Covers P1 (Source
 > Language is a Locale), P2 (Locale Identifiers Standardized), P3
 > (Resources External, Not Inline), P4 (Plural/Gender Parameterized),
 > and P10 (Translation Reversible and Versioned). Referenced by
 > `formatting.md` (strings the formatter resolves) and
 > `rtl-bidi.md` (the `dir` the resource layer carries).
 ## Resources are the Boundary (P3 Resources are External, Not Inline)
 - User-facing strings live in locale resource files, addressed by
  key. Code references a key; the resource layer resolves the key to
  the active locale. The source language is itself a locale
  (`en-US`), not a fallback baked into code.
 - String concatenation in code (`"Welcome, " + name + "!"`) is the
  cardinal violation: it bakes in source-language word order and
  breaks for every locale with different grammar. Replace every
  concatenation with a parameterized message:
  `t("welcome", { name })`.
 - The resource is the single place a string lives. Editing a string
  in code instead of the resource is a locality violation (C4): the
  string and its locale consequences now live apart.
 ## Resource File Formats
 | Format | Shape | When | Notes |
 |--------|-------|------|-------|
 | `.po` / `.pot` | gettext; msgid → msgstr, plural headers | Server-side, GNU ecosystem, PHP/Python/C | Mature tooling (`xgettext`, `msgmerge`); supports plural categories via header |
 | JSON (flat or namespaced) | `{ "key": "value" }` per locale | JS/web, i18next, FormatJS | Simple, machine-readable, but no native plural support — wrap with ICU MessageFormat |
 | Fluent `.ftl` | Mozilla FTL; asymmetric, resolver-driven | Browser-grade l10n, asymmetric translations | One message can resolve differently per locale without code changes; supports attributes, selectors |
 | ICU Resource Bundle | ICU binary/text resources | ICU-native, JVM, C++ | Tightest integration with ICU formatting/CLDR; steeper tooling |
 - None is advocated over the others. The choice is ecosystem fit,
  not correctness. All four satisfy P3/P4 when used as the boundary.
 - A custom format (a hand-rolled `.csv` of strings) is a violation:
  it is unsupported by standard tooling, has no plural grammar, and
  cannot compose with `formatting.md`'s ICU layer.
 ## Key Naming and Namespaces (P2 Locale Identifiers Standardized)
 - Locale identifiers are BCP 47 tags (`en-US`, `ar-EG`, `zh-Hans-CN`).
  No ad-hoc codes. The resource file is named for its locale:
  `en-US.json`, `ar-EG.po`, `ftl/ar-EG/main.ftl`.
 - Message keys are stable, semantic, and structured — not prose.
  `checkout.cart.item_count` not `"You have 3 items in your cart"`.
  A key that is the source string (`t("You have items")`) breaks the
  moment the source copy is edited; the key must outlive the copy.
 - Namespaces segment by surface (`checkout.*`, `errors.*`, `onboarding.*`)
  so that a locale can be loaded incrementally and so that key
  collisions across surfaces are impossible. A flat namespace with
  thousands of keys is a C2 (Clarity) violation waiting to happen.
 ## Fallback Chains
 - The fallback chain is explicit: requested locale → language-only
  (`en` from `en-GB`) → default locale → key itself (last resort).
  The default locale is declared once, not re-derived in every call
  site.
 - A missing key in the requested locale falling back silently to the
  source locale is a P3 violation: the user is silently shown the
  developer's locale, which is not the locale they asked for. Missing
  keys must be observable (see `testing-i18n.md`).
 - Fallback is a property of the resource layer, not of individual
  components. A component that re-implements fallback is duplicating
  a contract (C6 Composability violation).
 ## Plural and Gender in Resources (P4 Plural/Gender Parameterized)
 - Plural variants live in the resource, selected by the formatter,
  parameterized by the count. The code passes the count; the resource
  carries the variants; the formatter picks the right one per the
  locale's CLDR plural rules.
  ```
  // JSON + ICU MessageFormat (FormatJS / i18next)
  {
    "cart.item_count": "{count, plural, one {# item} other {# items}}"
  }
  // ar-EG.json — six plural categories per CLDR
  {
    "cart.item_count": "{count, plural, zero {لا عناصر} one {عنصر واحد} two {عنصران} few {# عناصر} many {# عنصرًا} other {# عنصر}}"
  }
  ```
 - `if (n == 1)` branching in code is a violation regardless of
  language. Arabic has six plural categories; Russian has three;
  English has two. A two-branch `if` encodes exactly one locale's
  rules and is wrong for every other.
 ## Extraction Tooling (P1, P3)
 - Strings are extracted mechanically (e.g. `xgettext`, `i18next-
  parser`, FormatJS babel plugin), not by hand-tagging. Mechanical
  extraction produces a `.pot` template that translators work from;
  the template is regenerated on every build.
 - A string that cannot be extracted (built at runtime from
  fragments) is a P3 violation: it is invisible to the pipeline. If
  the extractor cannot see it, neither can the translator.
 - The extracted template is versioned (`P10`): the diff between
  templates is the change in translatable surface. A template that
  is not committed is a contract that is not reviewable.
 ## Versioning and Rollback (P10 Translation Reversible and Versioned)
 - Resource files are committed to git alongside code. A bad
  translation is a `git revert` of the resource, not a hot-patch over
  the translator's work. Every locale resource has history,
  provenance (which translator / which service produced which
  commit), and a rollback path.
 - A locale resource that is generated by a translation service and
  committed without review is a P10 violation: the resource is
  versioned but the provenance is opaque. Review the diff the same
  way you review a code diff.
 ## What Violates Locale-Resource Discipline
 | Violation | Principle |
 |-----------|-----------|
 | `t("You have " + n + " items")` concatenation | P3 Resources are External, Not Inline |
 | A custom `.csv` string store instead of a standard format | P3 Resources are External, Not Inline |
 | Locale file named `english.json` not `en-US.json` | P2 Locale Identifiers are Standardized |
 | `if (n == 1) { t("item") } else { t("items") }` in code | P4 Plural and Gender are Parameterized |
 | A key equal to the source string (`t("Welcome back")`) | P2 / P10 — keys must outlive copy |
 | Silent fallback to the source locale with no signal | P3 Resources are External, Not Inline |
 | A runtime-built string the extractor cannot see | P3 Resources are External, Not Inline |
 | Resource files committed by a bot with no human review | P10 Translation is Reversible and Versioned |
 ## Relationship to Other Domains
 - `domains/uiux/copywriting.md` — copy lives in resources; UI
  microcopy is the source content the resource layer carries.
 - `domains/api/error-responses.md` — API error messages are locale-
  resource keys resolved at the boundary, not inline strings.
 - `domains/data/schema-design.md` — locale identifiers and resource
  shapes are a data contract; treat them as schema.
 - `domains/i18n/formatting.md` — the formatter resolves the
  parameterized message this layer produces.
@@ -0,0 +1,117 @@
 # RTL and Bidi — Derived Rules
 > Derives from `domains/i18n/first-principles.md`. Covers P6 (Text
 > Direction is a Layout Primitive) and P7 (Layout Accommodates
 > Expansion). Referenced by `testing-i18n.md` (RTL coverage is an
 > e2e tier). Grounded in W3C i18n bidi authoring, UAX #9, and
 > `domains/uiux/accessibility.md`.
 ## Text Direction is a Layout Primitive (P6 Text Direction is a Layout Primitive)
 - RTL and bidi are first-class layout concerns, not a CSS
  afterthought. The layout is designed for both directions from the
  first commit, not retrofitted when an RTL locale ships.
 - Logical CSS properties over physical properties, always. The
  browser resolves logical → physical from the `dir` attribute; the
  code never has to.
  | Physical (LTR-only) | Logical (dir-aware) | Resolves to in RTL |
  |---------------------|---------------------|--------------------|
  | `margin-left` | `margin-inline-start` | `margin-right` |
  | `margin-right` | `margin-inline-end` | `margin-left` |
  | `padding-left` | `padding-inline-start` | `padding-right` |
  | `left: 0` | `inset-inline-start: 0` | `right: 0` |
  | `text-align: left` | `text-align: start` | `text-align: right` |
  | `float: left` | use flexbox/grid + `inline-start` where supported | mirrored |
 - The `dir` attribute is set on the document root (`<html dir="rtl">`)
  and on subtrees whose direction differs from the document
  (`<span dir="ltr">` for an embedded Latin run). `dir` is the
  contract the bidi algorithm (UAX #9) reads; do not fake direction
  with `text-align` alone.
 ## The Bidi Algorithm (UAX #9)
 - The Unicode bidi algorithm resolves inline reordering of mixed-
  direction runs. The browser applies it; the author's job is to
  mark direction correctly, not to reorder by hand.
 - A string like `"The price is 15 USD"` in an RTL context renders
  with the Latin run `"15 USD"` in LTR within the RTL line — the
  algorithm handles it *if* the container's `dir` is set. Without
  `dir`, numbers and Latin fragments drift to the wrong edge.
 - `dir="auto"` on a container infers direction from the first strong
  directional character of its content — useful for user-generated
  content whose direction is unknown. `dir="auto"` is not a
  replacement for `dir="rtl"` on a known-RTL document.
 ## Mirroring (Icons, Controls, Diagrams)
 - Direction-aware icons mirror in RTL: a "back" arrow pointing left
  in LTR points right in RTL. A "refresh" circular arrow does not
  mirror. The rule: icons that imply direction mirror; icons that
  imply time or rotation do not.
 - Use `[dir="rtl"]` selectors or logical icon variants — never
  `transform: scaleX(-1)` as a one-off hack scattered across
  components. Centralize the mirroring rule (a token, a component
  prop) so it is auditable.
 - Numbers do not mirror. `"15 USD"` in an RTL line is still
  `"15 USD"` left-to-right inside the bidi run; mirroring it to
  `"DSU 51"` is a correctness violation.
 - Diagrams and flowcharts: a left-to-right process flow in LTR is a
  right-to-left flow in RTL. Decide per diagram whether the flow
  mirrors (most do) or is direction-neutral (some scientific
  schematics).
 ## Layout Accommodates Expansion (P7 Layout Accommodates Expansion)
 - Translated text expands. German is ~30% longer than English;
  Japanese is often shorter but taller; RTL mirroring shifts every
  visual anchor. Layouts are flexible:
  - No fixed pixel widths on translatable text containers.
  - No `white-space: nowrap` on translatable strings.
  - No `text-overflow: ellipsis` without a `title` carrying the full
    string.
  - Buttons sized to fit their longest locale variant, not the
    source.
 - A layout that breaks at +30% width is a layout that is wrong for
  most of the world's locales. Designing for the worst case up front
  is cheaper than reworking every screen when the first long-form
  locale ships.
 ## Common Pitfalls
 | Pitfall | Why it breaks | Fix |
 |---------|---------------|-----|
 | `margin-left` everywhere | In RTL the start is the right; `margin-left` leaves the right side unstyled | `margin-inline-start` |
 | `text-align: left` for "default" alignment | In RTL the default is right; `left` pins content to the wrong edge | `text-align: start` |
 | Icons hardcoded to LTR orientation | "Back" arrow points the wrong way in RTL | Mirror direction-implying icons via `[dir="rtl"]` |
 | Numbers mirrored with the layout | Numbers are LTR inside RTL; mirroring produces garbage | Leave number runs LTR; the bidi algorithm handles embedding |
 | Fixed `width: 120px` on a button | German button label overflows and truncates | `min-width` + `max-width` + flex; let content size |
 | `position: absolute; left: 0` | Pins to the physical left in both directions | `inset-inline-start: 0` |
 | Fake direction with `text-align` only | The bidi algorithm reads `dir`, not `text-align`; mixed runs reorder wrong | Set `dir` on the container |
 ## What Violates RTL/Bidi Discipline
 | Violation | Principle |
 |-----------|-----------|
 | A layout with no `dir` attribute, assuming LTR | P6 Text Direction is a Layout Primitive |
 | `margin-left` / `left: 0` / `text-align: left` throughout | P6 Text Direction is a Layout Primitive |
 | A "back" arrow that points left in the RTL build | P6 Text Direction is a Layout Primitive |
 | Numbers mirrored to read right-to-left | P6 Text Direction is a Layout Primitive |
 | `width: 100px` on a text container that overflows in German | P7 Layout Accommodates Expansion |
 | `white-space: nowrap` on a translated label | P7 Layout Accommodates Expansion |
 | `dir` faked with `text-align` and no `dir` attribute | P6 Text Direction is a Layout Primitive |
 | No RTL build until the first RTL locale ships | P6 Text Direction is a Layout Primitive |
 ## Relationship to Other Domains
 - `domains/uiux/accessibility.md` — RTL support is an accessibility
  concern for non-Latin-script users; WCAG 2.1 AA requires that
  direction be set correctly.
 - `domains/uiux/components.md` — components are built with logical
  properties so they survive direction and expansion without per-
  locale overrides.
 - `domains/i18n/testing-i18n.md` — RTL coverage is an e2e-tier
  test; pseudo-locale mirroring surfaces direction bugs early.
 - `domains/i18n/locale-resources.md` — the `dir` is part of the
  locale's metadata, carried alongside the resource bundle.
@@ -0,0 +1,142 @@
 # Testing i18n — Derived Rules
 > Derives from `domains/i18n/first-principles.md`. Covers P8
 > (Pseudo-Locales Test Early) and the testing-discipline angle on
 > P3 (Resources External), P5 (Formatting Locale-Aware), P6 (Text
 > Direction), and P10 (Translation Versioned). Referenced by
 > `locale-resources.md` (missing-key detection) and `rtl-bidi.md`
 > (RTL coverage tier).
 ## Pseudo-Locales Test Early (P8 Pseudo-Locales Test Early)
 - A pseudo-locale is a synthetic locale that transforms the source
  strings to surface i18n defects before real translations arrive.
  Three transforms cover the three defect classes:
  | Pseudo-locale | Transform | Surfaces |
  |---------------|-----------|----------|
  | `en-XA` (accented) | `Wêlcômê tô thê çhêckôût` | Strings not extracted (raw source appears), encoding bugs |
  | `en-XB` (lengthened / "long") | `Wᴇʟᴄᴏᴍᴇ ᴛᴏ ᴛʜᴇ ᴄʜᴇᴄᴋᴏᴜᴛ──────` (~30% longer, bracketed) | Layout overflow, fixed widths, truncation |
  | `en-XC` (RTL-mirrored) | Source rendered with `dir="rtl"` and a Latin-in-RTL run | LTR-only layout assumptions, physical CSS properties |
 - Pseudo-locale tests are cheap: they run against source strings, no
  translator involved, no string freeze required. A failing pseudo-
  locale run is a bug found at the cheapest possible point in the
  pipeline. Finding the same bug after real translation is a C5
  (Reversibility) violation: the fix now costs a re-translation.
 ## Pseudo-Locale → Testing Pyramid Mapping (IDEATE-28)
 - The testing pyramid (`domains/testing/pyramid.md`) has three tiers;
  i18n tests map to each tier with a distinct signal. The mapping is
  deliberate: each tier catches a different class of defect, and
  skipping a tier leaves a blind spot.
  | Pyramid Tier | i18n Test | Defect Caught | Tooling Shape |
  |--------------|-----------|---------------|---------------|
  | **Unit** | Missing-key detection | A key referenced in code but absent from the resource bundle; a key present in the source locale but missing from a target locale | Static scan over the resource bundle + code AST; runs per file, no runtime |
  | **Integration** | Snapshot per locale | Formatted output for a fixture input differs across locales in a way that breaks the contract (wrong plural, wrong date, overflow) | Render a known fixture through the formatter per locale; snapshot-diff against the recorded baseline |
  | **e2e** | RTL coverage | The app renders and is navigable in `dir="rtl"`; no layout overflow, no off-screen controls, no LTR-pinned anchors | Browser-driven run against the `en-XC` pseudo-locale (or a real RTL locale); assert on layout, not just text |
 - Unit is the broad base (fast, runs on every commit), e2e is the
  narrow top (slow, runs on PR merge). Integration sits between.
  This mirrors `domains/testing/pyramid.md` exactly — i18n is not a
  special case; it is a domain that uses the same tiers.
 ## Unit Tier — Missing-Key Detection (P3 Resources External)
 - A static scan compares the set of keys referenced in code against
  the keys present in each locale bundle. A key in code but not in
  `en-US` is a P3 violation (the string is not in the resource
  layer). A key in `en-US` but not in `ar-EG` is a coverage gap —
  the missing-key scan flags it before the locale ships.
 - Missing keys fail the build, not the runtime. A missing key that
  surfaces only when a user switches locale is a defect found in
  production, which is the most expensive place to find it.
  ```
  // tool output (illustrative)
  // missing-key scan
  [FAIL]  ar-EG: key "checkout.cart.item_count" referenced in code,
          absent from ar-EG.json
  [FAIL]  en-US: key "checkout.cart.total" referenced in Checkout.tsx:42,
          absent from en-US.json (not extracted)
  [PASS]  en-US, ar-EG, de-DE, zh-Hans-CN: all other keys present
  ```
 ## Integration Tier — Snapshot per Locale (P5 Formatting Locale-Aware)
 - For a fixed fixture input, render the formatted output per locale
  and snapshot it. A change in the snapshot is either an intended
  change (new CLDR data, new copy) or a regression.
 - The snapshot is per locale, not per format string. The same
  fixture (`{ count: 1, currency: "EUR", date: 2024-11-07 }`)
  produces different snapshots for `en-US`, `de-DE`, `ar-EG` — and
  that difference is the assertion. A locale whose snapshot matches
  the source locale's is a red flag: the formatter is not actually
  locale-aware.
  ```
  // snapshot — checkout.cart (fixture: count=1, currency=EUR, date=2024-11-07)
  // en-US
  "1 item · €1,234.56 · 11/7/2024"
  // de-DE
  "1 Artikel · 1.234,56 € · 07.11.2024"
  // ar-EG
  "عنصر واحد · ١٬٢٣٤٫٥٦ € · ٧/١١/٢٠٢٤"
  ```
 - Snapshots are reviewed, not rubber-stamped. A snapshot diff that
  changes the plural form for `ar-EG` is either a CLDR update (verify)
  or a regression (revert).
 ## e2e Tier — RTL Coverage (P6 Text Direction is a Layout Primitive)
 - A browser-driven run against `en-XC` (or a real RTL locale like
  `ar-EG`) asserts that the app is navigable in RTL: no overflow, no
  off-screen controls, no LTR-pinned anchors. The assertion is on
  layout, not on text — text correctness is the integration tier's
  job.
 - RTL e2e is the narrow top of the i18n pyramid: it is slow, it
  requires a browser, and it catches the defects the lower tiers
  cannot (the interaction of `dir` with the real layout engine). It
  runs on PR merge, not on every commit.
 ## Snapshot Discipline (P10 Translation Reversible and Versioned)
 - Snapshots are versioned in git. A snapshot that changes because of
  a real translation update is a committed diff, reviewed like a
  code change. A snapshot that changes because of a regression is a
  `git revert`.
 - A snapshot that is regenerated and committed without review is a
  P10 violation: the snapshot is versioned but the provenance is
  opaque. The same discipline applies to snapshots as to resources
  (`locale-resources.md`).
 ## What Violates i18n Testing Discipline
 | Violation | Principle |
 |-----------|-----------|
 | First i18n test runs against real translations, not pseudo-locales | P8 Pseudo-Locales Test Early |
 | No missing-key scan — gaps surface only at runtime in production | P3 Resources are External, Not Inline |
 | Snapshot per locale that matches the source locale's snapshot | P5 Formatting is Locale-Aware |
 | No RTL e2e — "we'll test RTL when we ship an RTL locale" | P6 Text Direction is a Layout Primitive |
 | Snapshots regenerated and committed without review | P10 Translation is Reversible and Versioned |
 | i18n tests only at e2e (no unit/integration tier) | pyramid inversion — `domains/testing/pyramid.md` |
 | Pseudo-locale run skipped because "it's not a real locale" | P8 Pseudo-Locales Test Early |
 ## Relationship to Other Domains
 - `domains/testing/pyramid.md` — the pseudo-locale → pyramid mapping
  mirrors this domain's unit / integration / e2e tiers exactly.
 - `domains/testing/fixtures.md` — locale fixtures (a fixed input
  rendered per locale) are the fixture shape for the integration
  tier.
 - `domains/i18n/locale-resources.md` — missing-key detection is the
  unit-tier scan over the resource bundle this doc defines.
 - `domains/i18n/formatting.md` — the integration-tier snapshot
  asserts against the formatter's output.
 - `domains/i18n/rtl-bidi.md` — the e2e tier exercises the layout
  rules this doc establishes.
 - `domains/uiux/accessibility.md` — RTL coverage is an a11y
  concern; an untested RTL build is an untested a11y surface.
@@ -0,0 +1,379 @@
 # Delivery Semantics — Derived Rules
 > Derives from `domains/messaging/first-principles.md`. Applies P4
 > (Delivery Semantics are Explicit), P3 (Consumers are Idempotent),
 > P2 (Ordering is a Property, Not an Assumption), P5 (Dead-Letter
 > Handling is Defined), and P6 (Backpressure is Bounded) primarily,
 > with P10 (DLQ depth as an alert). For the dead-letter strategy
 > decision, see the comparison table below. Includes a fenced
 > idempotency-key dedup-store example (IDEATE-39). Cross-links
 > `domains/concurrency/patterns` for the in-process retry/backoff
 > analog, `domains/errors/patterns` for errors as data for message
 > failures, and `domains/observability/metrics` for DLQ depth as an
 > alert.
 ## The Three Delivery Semantics (P4 Delivery Semantics are Explicit)
 - **At-most-once**: a message is delivered 0 or 1 times; loss is
  possible, duplication is not. The producer fires-and-forgets; the
  broker does not ack; the consumer does not dedup. Lowest latency,
  lowest implementation cost, lossy. Fits telemetry where a dropped
  sample is acceptable (MQTT QoS 0, fire-and-forget metrics).
 - **At-least-once**: a message is delivered 1 or more times;
  duplication is possible, loss is not. The producer sends and
  waits for a broker ack; the consumer processes and acks; a crash
  before the consumer's ack triggers redelivery. The consumer must
  be idempotent (P3). The default for side-effecting operations
  (orders, payments, commands). The vast majority of broker-backed
  queues (SQS standard, RabbitMQ ack, MQTT QoS 1).
 - **Exactly-once**: a message is delivered exactly 1 time; no
  loss, no duplication. In practice, this is at-least-once plus
  idempotency (P3), or a transactional consume-process-produce loop
  (P4 — see `domains/messaging/streams.md`). Jepsen analyses verify
  broker claims: "exactly-once" requires independent verification;
  the durable engineering practice is at-least-once with idempotent
  consumers (P3), which collapses to exactly-once under correct
  dedup.
 - The semantic is declared per channel (P4), not emergent. An
  unstated semantic is a defect: the consumer guesses, and the
  guess is wrong under the first failure. See
  `domains/messaging/queues.md` for the three-semantics comparison
  table (semantics, latency cost, implementation cost, when each
  fits).
 ## Idempotency (P3 Consumers are Idempotent)
 - Idempotency is the correctness property that makes at-least-once
  safe. A consumer that processes the same message twice has the
  same effect as processing it once. The mechanism is the
  idempotency key: a per-message unique identifier the consumer
  uses to dedup redeliveries.
 - The idempotency key is per-message, not per-producer or
  per-session. A consumer that dedups by producer alone drops
  distinct messages issued in the same window. Use a UUID per
  message, or a deterministic key derived from the message content
  (e.g., `(entity, operation, version)`).
 - The dedup store is bounded (P6 — Backpressure is Bounded): a
  dedup store that grows without bound is a memory leak. Use a TTL
  window longer than the broker's max-redelivery window, or a
  bounded LRU. The TTL is the P6 bound: a key seen within the TTL
  is a redelivery; a key older than the TTL is expired (the broker
  has given up redelivering it).
 ## Idempotency-Key Dedup Store (IDEATE-39, P3 — fenced example)
 The dedup store is the concrete mechanism that makes a consumer
 idempotent under at-least-once delivery. This is the fenced
 example required by IDEATE-39 (parallel to the v0.3 IDEATE-29
 signed-attestation fenced example): it is NOT prose-only — the
 consumer-with-dedup-store demonstrates P3 concretely.
 ```python
 # Idempotent consumer with a dedup store (P3 Consumers are
 # Idempotent, P6 Backpressure is Bounded — the dedup store is
 # TTL-bounded). The consumer dedups by idempotency key before
 # processing; a redelivered message is a no-op, not a double-apply.
 import time
 # P6: the dedup store is bounded by a TTL window. The TTL must
 # exceed the broker's max-redelivery window; beyond the TTL, the
 # key is expired (the broker has given up). A dedup store with no
 # TTL is a memory leak (P6 violation).
 DEDUP_TTL_SECONDS = 24 * 3600   # longer than max-redelivery window
 class DedupStore:
    """A TTL-bounded idempotency-key dedup store (P3, P6).
    seen(key): True if the key was processed within the TTL.
    mark(key): Record the key as processed (with a timestamp).
    """
    def __init__(self, backend):
        # backend is a Redis, a DB, or an in-process LRU. The
        # backend must be shared across consumer instances if the
        # subscription is shared (P2/P3 — see domains/messaging/
        # pubsub.md on shared vs independent subscriptions).
        self.backend = backend
    def seen(self, key: str) -> bool:
        ts = self.backend.get(key)
        if ts is None:
            return False
        if time.time() - ts > DEDUP_TTL_SECONDS:
            # P6: expired. The broker has given up redelivering;
            # this key is no longer a redelivery signal.
            self.backend.delete(key)
            return False
        return True
    def mark(self, key: str):
        self.backend.set(key, time.time(), ttl=DEDUP_TTL_SECONDS)
 # The idempotent consumer: dedup before process, mark after
 # process, ack after mark. A crash before mark re-processes (the
 # dedup store does not have the key); a crash before ack
 # redelivers (the broker did not see the ack) and the dedup store
 # makes the redelivery a no-op (P3).
 def consume_idempotent(broker, dedup: DedupStore, process):
    for message in broker.receive():
        # P3: dedup BEFORE process. A redelivered message is a
        # no-op, not a double-apply.
        if dedup.seen(message["idempotencyKey"]):
            broker.ack(message)        # already processed; skip
            continue
        try:
            process(message["payload"])
            # P3: mark AFTER process succeeds. A crash between
            # process and mark re-processes (acceptable: the
            # process must be idempotent OR the mark must be
            # transactional with the process — see below).
            dedup.mark(message["idempotencyKey"])
            broker.ack(message)
        except TransientError as exc:
            # P6: bounded retry with backoff. Nack for redelivery;
            # the broker redelivers after exponential backoff.
            broker.nack(message, delay=backoff(message["attempt"]))
        except PoisonError as exc:
            # P5: poison message. Route to DLQ, do NOT retry
            # forever (see the DLQ routing rule below).
            route_to_dlq(broker, message, exc)
            broker.ack(message)       # remove from the origin queue
 ```
 - The order `process → mark → ack` gives at-least-once with
  idempotent dedup: a crash before `mark` re-processes (the dedup
  store does not have the key), and a crash before `ack`
  redelivers (the dedup store makes the redelivery a no-op). If
  `process` is not itself idempotent, the `process → mark` window
  must be transactional (e.g., process and mark in one DB
  transaction) — otherwise a crash in the window double-applies.
 - For a non-idempotent `process` (e.g., a payment that must not
  double-charge), use a transactional dedup: process and mark in
  one DB transaction, so the mark commits iff the process
  commits. This is the "exactly-once via idempotency" pattern
  (P4): at-least-once delivery plus a transactional
  process-and-mark collapses to exactly-once under correct
  transactional semantics.
 ## Ordering (P2 Ordering is a Property, Not an Assumption)
 - The delivery semantic interacts with ordering (P2). At-least-once
  with per-partition ordering: a redelivery within a partition
  preserves order (the redelivered message re-appears in its
  original position relative to other messages the consumer has
  not yet seen). At-least-once with no ordering: a redelivery may
  appear out of order relative to messages delivered after it.
 - The consumer must not assume an ordering property the broker
  does not provide (P2). A standard queue delivers per-receive-node
  arrival order but no global order and no order across redeliveries;
  a FIFO queue delivers strict per-group order including across
  redeliveries; a partitioned stream delivers strict per-partition
  order across redeliveries. Document the property; do not assume
  it. See `domains/messaging/queues.md` (FIFO vs standard) and
  `domains/messaging/streams.md` (per-partition order).
 ## Dead-Letter Strategies (P5 Dead-Letter Handling is Defined)
 - A poison message (unparseable, repeatedly failing, or exhausting
  the retry budget) must be routed to a dead-letter queue, not
  retried forever or silently dropped (P5). The DLQ is observable
  (P10 — depth is an alert) and drainable (an operator can
  inspect, replay, or discard with audit).
 - The dead-letter strategy determines when a message is
  dead-lettered and what the operator sees. The choice is the
  decision matrix below (D-069).
 ## Dead-Letter Strategy Comparison (D-069)
 | Strategy | When It Applies | Failure Visibility | Operational Cost |
 |----------|-----------------|-------------------|------------------|
 | **Retry-count-limit** | A fixed max-redeliveries count (e.g., 5). After N redeliveries, route to DLQ. Simple, predictable. | The redelivery count is visible in the DLQ entry; the operator sees how many times it was retried. | Low — a counter per message; no backoff tuning. Risk: retries fire as fast as the broker redelivers, hammering a downstream that is already failing (P6 — no backoff = no backpressure escape). |
 | **TTL-with-backoff** | A max time-to-live for redelivery (e.g., 30 minutes) with exponential backoff between retries. After the TTL, route to DLQ. | The TTL and the backoff schedule are visible; the operator sees the retry timeline. | Medium — backoff tuning per message type. Benefit: backoff gives the downstream time to recover (P6 — the retry rate is bounded); fits transient failures (a downstream that is briefly unavailable). |
 | **Poison-queue** | A separate queue for messages that fail a specific check (unparseable, schema-invalid, unknown type) before any processing retry. Routed immediately, not retried. | The poison queue is a separate signal from the DLQ; the operator sees parse-vs-process failures distinctly. | Low — a routing rule per check. Benefit: distinguishes "never going to succeed" (poison) from "might succeed on retry" (DLQ). Use for unparseable messages that no retry will fix. |
 | **DLQ + alert** | Any of the above strategies, plus an alert on DLQ depth. The DLQ is observable (P10) — depth, age, and rate are alerted. | Highest — the operator is paged on DLQ growth; the DLQ is a first-class signal, not a graveyard. | Medium — alerting setup per DLQ. This is the P5/P10 floor: a DLQ without an alert is a silent correctness defect (poison messages accumulate invisibly). |
 - The default for transient failures is **TTL-with-backoff + DLQ
  + alert**: backoff gives the downstream time to recover (P6),
  the TTL bounds the retry budget (P5), the DLQ captures the
  unprocessable, and the alert makes it visible (P10). The default
  for unparseable messages is **poison-queue + alert**: route
  immediately, do not retry a message no retry will fix.
 - **Retry-count-limit alone** (no backoff, no alert) is the
  `messaging-unbounded-retry` chaos anti-pattern's cousin: it
  caps the count but hammers the downstream at full retry rate,
  and the DLQ grows silently if no alert is wired. Always pair
  a retry budget with backoff (P6) and an alert (P10).
 - The failure-visibility column is the P10 check: every strategy
  must surface the failure to the operator. A strategy with no
  visibility is a P10 violation regardless of its retry semantics.
 - The operational-cost column is the C8 tradeoff: more visibility
  and more backoff cost more to set up but pay back in operational
  stability. The P5/P10 floor is "DLQ + alert"; below that, the
  strategy is a silent defect waiting to grow.
 ## DLQ Routing Rule (P5, P10)
 ```python
 # DLQ routing rule (P5 dead-letter handling, P6 bounded retry with
 # backoff, P10 DLQ depth alert). Combines TTL-with-backoff for
 # transient failures and poison-queue for unparseable messages,
 # with an alert on DLQ depth.
 import json, time
 MAX_RETRY_TTL_SECONDS = 30 * 60    # 30 min total retry window
 DLQ = "orders-dlq"
 POISON = "orders-poison"           # unparseable; never retried
 def route_to_dlq(broker, message, reason):
    """Route a message to the DLQ with audit metadata (P5)."""
    broker.send(DLQ, body=json.dumps({
        "original": message,
        "reason": str(reason),
        "deadLetteredAt": now_iso(),
        "redeliveryCount": message.get("attempt", 0),
    }))
    # P10: emit a metric so DLQ depth alerts fire. A DLQ that
    # grows with no alert is a silent correctness defect (P5/P10).
    metrics.increment("dlq.depth", tags={"queue": "orders"})
    # The ack removes the message from the origin queue; the DLQ
    # is the durable record (P5 — observable and drainable).
 def consume_with_dlq(broker, dedup, process):
    for message in broker.receive():
        # P1: parse first. An unparseable message is poison —
        # route immediately, do NOT retry (no retry will fix it).
        try:
            payload = json.loads(message["body"])
        except (ValueError, SchemaError) as exc:
            broker.send(POISON, body=json.dumps({
                "original": message["body"],
                "reason": f"parse-failed: {exc}",
                "poisonedAt": now_iso(),
            }))
            metrics.increment("poison.depth", tags={"queue": "orders"})
            broker.ack(message)       # remove from origin; poison queue holds it
            continue
        # P3: dedup before process.
        if dedup.seen(payload["idempotencyKey"]):
            broker.ack(message); continue
        attempt = payload.get("attempt", 0)
        first_attempt_ts = payload.get("firstAttemptTs", time.time())
        try:
            process(payload)
            dedup.mark(payload["idempotencyKey"])
            broker.ack(message)
        except TransientError as exc:
            # P6: TTL-with-backoff. If the retry window is
            # exhausted, route to DLQ; otherwise redeliver with
            # exponential backoff.
            if time.time() - first_attempt_ts > MAX_RETRY_TTL_SECONDS:
                route_to_dlq(broker, message, exc)   # P5
                broker.ack(message)
            else:
                broker.nack(message, delay=backoff(attempt))
        except PermanentError as exc:
            # A permanent error (e.g., a not-found dependency)
            # does not benefit from retry — route to DLQ now.
            route_to_dlq(broker, message, exc)
            broker.ack(message)
 ```
 - The routing rule distinguishes three failure modes: **poison**
  (unparseable — route immediately, no retry), **transient**
  (retry with backoff until the TTL, then DLQ), and **permanent**
  (a retry will not fix it — DLQ now). This distinction is the P5
  discipline: not every failure is a retry; some are immediate
  DLQs.
 - The DLQ entry carries `reason`, `deadLetteredAt`, and
  `redeliveryCount` — it is auditable (the operator knows why each
  message was dead-lettered and how many times it was retried).
  This is the errors-as-data discipline — see
  `domains/errors/patterns` for the general principle a DLQ entry
  instantiates.
 ## Retry Budgets and Backoff (P6 Backpressure is Bounded)
 - The retry budget is the cap on redelivery: a count, a TTL, or
  both. After the budget, the message routes to the DLQ (P5). An
  unbounded retry budget is the `messaging-unbounded-retry` chaos
  anti-pattern (P5 breach): the consumer never makes progress past
  the poison message.
 - **Exponential backoff** spaces retries: 1s, 2s, 4s, 8s, ... with
  a jitter to avoid thundering-herd synchrony. Backoff gives the
  downstream time to recover (P6 — the retry rate is bounded,
  giving the downstream a chance to catch up). A retry with no
  backoff hammers the downstream at full rate, making the failure
  worse.
 - The retry budget × the backoff schedule is the P6 bound: the
  consumer's retry load is bounded by design, not by luck. See
  `domains/concurrency/patterns` Pattern 6 (Timeout on Every
  Block) for the in-process retry/backoff analog; messaging owns
  the broker-backed instance where the redelivery comes from the
  broker across a network, not an in-process loop (D-062).
 ## Poison Messages (P5, P1)
 - A poison message is one no retry will fix: unparseable (the
  schema is wrong, P1), unknown type (the consumer does not handle
  this version, P9), or a permanent failure (a not-found
  dependency). Retrying a poison message wastes resources and
  blocks the queue (P6 — the consumer never makes progress).
 - Poison messages route to the poison queue immediately (no
  retry), distinct from the DLQ (which holds messages that
  exhausted their retry budget on transient failures). The
  distinction is the P5 discipline: a poison queue is for
  "never going to succeed"; a DLQ is for "might have succeeded
  but didn't within the budget."
 - A poison queue without an alert is the same defect as a DLQ
  without an alert (P10): the operator cannot see the poison
  accumulating. Wire both to `domains/observability/metrics`.
 ## Cross-Link to Concurrency (P6, cross-link concurrency/patterns)
 - The retry/backoff discipline here is the cross-process analog
  of `domains/concurrency/patterns` Pattern 6 (Timeout on Every
  Block) for in-process retry. Concurrency owns the in-process
  analog (a retry loop within one program, with a timeout per
  attempt); messaging owns the broker-backed instance (the broker
  redelivers across a network, the consumer applies backoff via
  nack-with-delay). The failure model differs: in-process retry
  fails by a thread crash or a timeout; broker-backed retry fails
  by network partition, broker restart, or consumer crash-and-retry
  (D-062).
 - The cross-link is one-directional outward (messaging →
  concurrency) per D-026 extended: messaging references concurrency
  as the in-process foundation; concurrency does not back-link to
  messaging.
 ## Cross-Link to Errors (P5, cross-link errors/patterns)
 - A poison message is an errors-as-data instance: the failure is
  captured as a DLQ entry (with `reason`, `redeliveryCount`,
  `deadLetteredAt`), not swallowed. See `domains/errors/patterns`
  for the general errors-as-data discipline a DLQ entry
  instantiates. The DLQ is the async-messaging instance of an
  error log — observable, auditable, drainable.
 - The cross-link is one-directional outward (messaging → errors):
  messaging references errors for the errors-as-data pattern;
  errors does not back-link to messaging.
 ## What Violates Delivery-Semantics Discipline
 | Violation | Principle |
 |-----------|-----------|
 | Unstated delivery semantic (at-least-once vs exactly-once guessed) | P4 Delivery Semantics are Explicit |
 | Non-idempotent consumer under at-least-once delivery | P3 Consumers are Idempotent |
 | Dedup store with no TTL (memory leak; grows without bound) | P6 Backpressure is Bounded |
 | Dedup key per-producer (distinct messages in the same window deduped) | P3 Consumers are Idempotent |
 | Retry with no backoff (hammers the downstream at full rate) | P6 Backpressure is Bounded |
 | Unbounded retry budget (consumer never progresses past the poison) | P5 Dead-Letter Handling is Defined |
 | DLQ with no depth alert (poison messages accumulate invisibly) | P10, `domains/observability/metrics` |
 | Poison message retried forever (no poison queue, no immediate DLQ) | P5 Dead-Letter Handling is Defined |
 | Process-not-idempotent with non-transactional mark (crash in window double-applies) | P3, P4 |
 | DLQ entry with no reason/audit metadata (uninspectable failure) | P5, `domains/errors/patterns` |
 | Ordering assumption the broker does not provide (FIFO assumed on standard queue) | P2 Ordering is a Property, Not an Assumption |
@@ -0,0 +1,307 @@
 # Messaging — First Principles
 ## 1. The Principles
 ### P1. Messages are Contracts
 A message has an explicit, versioned schema. Producer and consumer
 agree on shape before exchange; the schema is the boundary, not a
 guess. A schemaless message — a free-form JSON blob the consumer
 parses by hope — is a defect: the consumer breaks silently on the
 next shape change, and the producer has no contract to evolve
 against. This derives from `C1 Correctness` (the exchange must
 carry what the parties agreed to) and `C2 Clarity` (the schema
 makes the boundary obvious to both sides). This is the
 cross-process expression of the contract discipline that
 `domains/api/rest` owns for synchronous request/response: the
 message schema is to async exchange what the API contract is to
 sync exchange. It is distinct from
 `domains/concurrency/patterns` Pattern 1 (Message Passing), which
 owns the *in-process* channel primitive — here the contract spans
 separate systems and survives network failure (D-062). See
 `domains/messaging/queues.md` for the queue-flavored application
 and `domains/messaging/streams.md` for the durable-log-flavored
 application.
 ### P2. Ordering is a Property, Not an Assumption
 Ordering guarantees — per-partition strict, global, or none — are
 explicit and documented. "It's FIFO" is a claim that must be backed
 by the broker's partitioning contract, not an assumption the
 consumer makes and the broker may not honor. A standard queue
 delivers in arrival order per receive-node but offers no global
 ordering across shards; a FIFO queue delivers strict per-message-
 group order but at a latency cost; a partitioned stream delivers
 strict per-partition order but only within a partition. Each is a
 distinct, declared property. This derives from `C1 Correctness`
 (order is a correctness property — a consumer that assumes order
 the broker does not provide is wrong) and `C2 Clarity` (the
 ordering guarantee is documented, not discovered in production).
 This is distinct from in-process ordering, which
 `domains/concurrency/patterns` Pattern 1 owns for channels within
 one program: messaging ordering survives network failure, broker
 restart, and consumer crash-and-retry — a stronger failure model
 than thread-local channels (D-062). The `messaging-shared-
 subscription` chaos anti-pattern breaches this rule: two consumers
 sharing one subscription break per-consumer ordering because the
 broker dispatches each message to an arbitrary consumer. See
 `domains/messaging/streams.md` for the partition-order contract
 and `domains/messaging/delivery-semantics.md` for the interaction
 of ordering with the three delivery semantics.
 ### P3. Consumers are Idempotent
 Delivery is at-least-once by default across the network; a
 consumer deduplicates via idempotency keys or deterministic
 processing. "Exactly-once" is idempotency plus at-least-once, not a
 broker guarantee — Jepsen analyses of Kafka, RabbitMQ, and NATS
 establish that exactly-once claims require independent
 verification, and the durable engineering practice is to make
 consumers idempotent under redelivery. A non-idempotent consumer
 under at-least-once delivery doubles the effect on every retry; a
 non-idempotent consumer under a claimed exactly-once broker is a
 bug waiting for the broker's exactly-once invariant to break. This
 derives from `C1 Correctness`: correctness under redelivery is the
 contract, not a nice-to-have. This parallels
 `domains/edge/P5 Edge Operations are Idempotent` (the
 cross-partition device-and-cache-flavored analog) and is the
 cross-process instance of the retry-safety discipline that
 `domains/concurrency/patterns` Pattern 6 (Timeout on Every Block)
 implies for in-process retry. It is distinct from in-process
 retry because the redelivery comes from the broker across a
 network, not from an in-process loop (D-062). See
 `domains/messaging/delivery-semantics.md` for the idempotency-key
 dedup-store pattern.
 ### P4. Delivery Semantics are Explicit
 At-least-once / at-most-once / exactly-once is a declared choice
 per channel, not an emergent behavior. The tradeoff — latency cost,
 implementation complexity, operational cost — is made consciously
 and documented. At-most-once is fire-and-forget (low latency, lossy);
 at-least-once is acked with possible duplication (the default,
 requires idempotent consumers per P3); exactly-once is at-least-once
 plus idempotency or a transactional two-phase commit (highest cost,
 narrowest fit). An unstated semantic is a defect: the consumer
 guesses, and the guess is wrong under the first failure. This
 derives from `C1 Correctness` (the chosen semantic must hold) and
 `C2 Clarity` (the tradeoff is visible to the reader and the
 operator). This is the cross-process analog of the explicit-failure-
 mode discipline that `domains/errors/patterns` owns for synchronous
 code — messaging makes the delivery-mode choice as explicit as an
 error-handling choice. See `domains/messaging/queues.md` for the
 three-semantics comparison table and `domains/messaging/delivery-
 semantics.md` for the correctness properties of each.
 ### P5. Dead-Letter Handling is Defined
 Poison messages — unparseable, repeatedly failing, or exhausting
 the retry budget — are routed to a dead-letter queue, not retried
 forever or silently dropped. The DLQ is observable and drainable: an
 operator can inspect it, replay from it, or discard with audit. An
 unbounded retry loop is a livelock: the consumer never makes
 progress past the poison message. A silent drop is a correctness
 defect: the message vanished with no record. This derives from `C1
 Correctness` (poison messages must not livelock the consumer or
 silently disappear) and `C5 Reversibility` (the DLQ is the
 reversibility mechanism — a dead-lettered message can be reprocessed
 after the bug is fixed). This is the cross-process analog of the
 bounded-error discipline that `domains/errors/patterns` owns for
 synchronous code: a poison message is an error-as-data instance
 that must be observable and recoverable, not swallowed. It is
 distinct from in-process error handling because the failure spans a
 network and a consumer restart (D-062). See
 `domains/messaging/delivery-semantics.md` for the dead-letter
 strategy comparison table and the DLQ routing rule pattern.
 ### P6. Backpressure is Bounded
 A slow consumer cannot unbounded-buffer the broker or the
 producer. Backpressure is explicit: consumer lag is visible,
 max-unacked is bounded, the retry budget is capped. A consumer
 that falls behind without a visible signal is a silent backlog —
 the operator cannot fix what they cannot see, and the broker's
 memory grows without bound until it fails. This derives from `C1
 Correctness` (a backlog that grows until OOM is a correctness
 failure) and `C8 Economy` (the broker's memory is bounded by
 design, not by luck). This is distinct from
 `domains/concurrency/P9 Bounded Queues` and
 `domains/concurrency/patterns` Pattern 5 (Bounded Queue with
 Backpressure), which own the *in-process* analog: concurrency's
 bounded queue fails by OOM or thread crash; messaging's bounded
 backpressure fails by network partition, broker restart, or
 consumer crash-and-retry (D-062). The Reactive Streams
 specification (`request(n)`, `onNext` bounded) is the in-process
 instance; messaging's broker-backed backpressure is the
 cross-process instance above it. See `domains/messaging/queues.md`
 for prefetch and max-unacked and `domains/observability/metrics`
 for consumer-lag as an alert.
 ### P7. Partitioning is Intentional
 The partition key determines ordering, parallelism, and hotspots.
 Key choice is a design decision with documented rationale, not a
 default. A key that hashes unevenly creates a hot partition that
 limits throughput; a key that does not match the ordering need
 breaks per-key semantics; a key that is too coarse (one partition
 for the whole topic) serializes all the traffic. The partition
 count is a capacity bound: too few partitions cap parallelism, too
 many partition overhead the broker. This derives from `C4 Locality`
 (ordering and parallelism are co-located with the partition) and
 `C6 Composability` (the partition is the unit of parallelism and
 scaling — consumer groups compose from per-partition workers).
 This is the cross-process analog of the locality discipline that
 `domains/performance/` owns for generic data-near-compute
 optimization: performance's locality is algorithmic (data near
 compute); messaging's locality is partitional (order and
 parallelism near the partition). See `domains/messaging/streams.md`
 for the partitioned-log model and consumer-group rebalance
 strategies.
 ### P8. Replay and Retention are Configured
 Retention windows and replay-from-offset are explicit. A message
 is not ephemeral by default; the broker is a durable log, not a
 pipe. A topic with no retention is a fire-and-forget stream — a
 consumer that falls behind loses data permanently; a topic with
 infinite retention is an unbounded log — the broker grows until
 disk exhaustion. Both are defects: the retention window is a
 declared bound, and replay-from-offset is the mechanism that makes
 the log durable (re-consumable) rather than ephemeral. This derives
 from `C5 Reversibility` (a retained message is reversible — it can
 be re-consumed; an ephemeral message is not) and `C7 Observability`
 (the durable log is itself an observable record of what happened —
 the offset is the position from which to replay). This is the
 foundation for `domains/messaging/streams.md` and the rule that
 distinguishes a stream from a queue (a queue deletes on ack; a
 stream retains for replay). See `domains/messaging/pubsub.md` for
 the pub/sub-vs-stream durability boundary.
 ### P9. Schemas Evolve Compatibly
 Schema changes are backward- and forward-compatible by
 construction. Breaking changes are versioned migrations, not
 silent shape edits. A producer that ships a new field the old
 consumer ignores is backward-compatible; a consumer that handles a
 missing field the new producer omits is forward-compatible. A
 silent schema change — the producer renames a field and the
 consumer parses `undefined` — is a P1 violation (the contract was
 broken) compounded here as an evolution defect. This derives from
 `C5 Reversibility` (a schema change is reversible by versioning —
 the old shape is still readable) and `C6 Composability` (producers
 and consumers of different versions compose because the schema
 evolves compatibly). This parallels `domains/data/migrations`
 (schema migration for databases) and `domains/api/versioning`
 (API contract evolution): messaging's schema evolution is the
 async instance of the same compatibility discipline. See
 `domains/messaging/streams.md` for the stream-schema-evolution
 angle.
 ### P10. Messaging is Observable
 Consumer lag, DLQ depth, throughput, and consumer-group health are
 first-class signals. Silent backlog is a bug, not a feature: a
 consumer that falls behind with no lag metric is invisible until
 the downstream effect surfaces — by which time the backlog may be
 hours or days. A DLQ that grows without an alert is a silent
 correctness defect: poison messages are accumulating and no one
 knows. This derives from `C7 Observability` (the broker's behavior
 is visible to the operator) and `C1 Correctness` (backlog
 detection is a correctness bound — unbounded lag is a failure).
 This is distinct from `domains/observability/metrics`, which owns
 *generic* structured metrics; messaging owns the *broker-specific*
 signals — lag, DLQ depth, partition imbalance, consumer-group
 rebalance events. See `domains/observability/metrics` for the
 generic SLI/SLO discipline and `domains/observability/tracing` for
 cross-partition traces.
 ## 2. Core Principle Trace
 Each messaging P-rule derives from one or more core C-rules
 (C1–C8). The matrix extension lands in P4 of the v0.4 plan; the
 traces below are authoritative. Messaging is a broad-derivation
 domain touching 7 of 8 core principles (C1, C2, C4, C5, C6, C7,
 C8); C3 (Simplicity) is not a primary derivation — messaging is
 inherently a tradeoff domain where simplicity yields to the
 correctness of delivery guarantees (a simpler-than-necessary
 delivery model does not handle the failure cases, per C3's
 "simpler than necessary is also a violation").
 | P-rule | Core | Why |
 |--------|------|-----|
 | P1 Messages are Contracts | C1, C2 | Correctness of the exchange; clarity of the schema boundary |
 | P2 Ordering is a Property, Not an Assumption | C1, C2 | Correctness of order; clarity of the guarantee |
 | P3 Consumers are Idempotent | C1 | Correctness under redelivery |
 | P4 Delivery Semantics are Explicit | C1, C2 | Correctness of the chosen semantic; clarity of the tradeoff |
 | P5 Dead-Letter Handling is Defined | C1, C5 | Correctness of poison-message routing; reversibility of reprocessing |
 | P6 Backpressure is Bounded | C1, C8 | Correctness of bounded backlog; economy of broker memory |
 | P7 Partitioning is Intentional | C4, C6 | Locality of order; composability of parallelism |
 | P8 Replay and Retention are Configured | C5, C7 | Reversibility of replay; observability of the durable log |
 | P9 Schemas Evolve Compatibly | C5, C6 | Reversibility of schema changes; composability of versions |
 | P10 Messaging is Observable | C7, C1 | Observability of lag/DLQ; correctness of backlog detection |
 ## 3. What Violates These Principles
 | Violation | Principle Breached |
 |-----------|-------------------|
 | Schemaless message (no versioned contract; consumer parses by guess) | P1 Messages are Contracts |
 | "It's FIFO" with no documented partition contract | P2 Ordering is a Property, Not an Assumption |
 | Non-idempotent consumer under at-least-once delivery | P3 Consumers are Idempotent |
 | Unstated delivery semantic (at-least-once vs exactly-once guessed) | P4 Delivery Semantics are Explicit |
 | No dead-letter queue (poison message retried forever or silently dropped) | P5 Dead-Letter Handling is Defined |
 | Unbounded retry budget (no cap; slow consumer stalls the partition) | P6 Backpressure is Bounded |
 | Default partition key (no rationale; hotspot or wrong-order) | P7 Partitioning is Intentional |
 | Ephemeral broker (no retention; no replay) | P8 Replay and Retention are Configured |
 | Silent schema change (producer breaks consumers with no version bump) | P9 Schemas Evolve Compatibly |
 | Silent backlog (no lag metric; consumer falls behind invisibly) | P10 Messaging is Observable |
 | Shared subscription (two consumers share one subscription; per-consumer ordering breaks) | P2 Ordering is a Property, Not an Assumption (P3 compounding) |
 | Blocking consumer (slow downstream call with no timeout; broker redelivers to the stuck consumer) | P6 Backpressure is Bounded |
 ## 4. Relationship to Other Domains
 Messaging systems are the engineering discipline of
 **cross-process, cross-system asynchronous communication via
 brokers**. Producer and consumer are separate systems; the broker
 is the intermediary that brokers delivery, ordering, retention,
 and failure semantics. The distinguishing constraints are a
 cross-process failure model (network, not crash), explicit
 delivery semantics, decoupled producer/consumer lifecycle, and
 replay-and-retention as a durable-log property. Messaging overlaps
 `domains/concurrency/` by *subject* (messages, queues,
 backpressure) but not by *failure model*: per D-062, messaging
 owns the cross-process/network-failure-model angle; concurrency
 owns the in-process/crash-failure-model angle. The discriminator
 is the failure model: concurrency's queue fails by OOM or thread
 crash; messaging's queue fails by network partition, broker
 restart, or consumer crash-and-retry. Messaging extends
 concurrency's bounded-queue/backpressure model to the network-
 partition regime. Cross-links are one-directional outward (per
 D-026 extended); no back-link edits to v0.1/v0.2/v0.3 content.
 - `domains/concurrency/patterns` ← P6 (the broker-backed bounded
  queue is the cross-process analog of the in-process bounded
  buffer — concurrency Pattern 5 owns in-process; messaging owns
  the network-failure-model instance above it, per D-062)
 - `domains/concurrency/patterns` ← P3 (idempotent retry is the
  cross-process analog of in-process retry-safety — the failure
  model differs: broker redelivery across a network vs in-process
  loop)
 - `domains/observability/metrics` ← P10 (consumer lag and DLQ
  depth as alerts; observability owns the generic SLI/SLO
  discipline, messaging owns the broker-specific signals)
 - `domains/observability/tracing` ← P10 (cross-partition traces
  for stream processing; observability owns the generic tracing
  discipline, messaging owns the cross-partition propagation)
 - `domains/data/schema-design` ← P1, P9 (message schema design
  and evolution; data owns the generic schema discipline,
  messaging owns the cross-process message-shape instance)
 - `domains/errors/patterns` ← P5 (errors as data for message
  failures; a poison message is an error-as-data instance that must
  be observable and recoverable, not swallowed)
 - `domains/edge/iot` ← P4 (the edge↔messaging cross-link
  resolves bidirectionally here: edge/iot.md links outward to
  messaging/queues for MQTT QoS parallels to delivery semantics;
  this first-principles doc acknowledges the back-link — the
  edge/iot.md → messaging/queues link from P1 now resolves because
  messaging/queues.md exists, completing the bidirectionality per
  IDEATE-40)
 > Note: the edge/iot.md → messaging/queues cross-link (MQTT QoS
 > parallels for delivery semantics) was authored in P1 with a
 > dangling reference; this P2 authorship of messaging/queues.md
 > resolves it. The bidirectionality is verified in P5
 > (ATELIER-114 per IDEATE-40). The cross-link is one-directional
 > outward from edge/iot.md; this first-principles doc
 > acknowledges the resolution without editing edge/iot.md (per
 > D-026 extended — no back-link edits to v0.1/v0.2/v0.3 or to
 > P1-authored edge content).
@@ -0,0 +1,269 @@
 # Pub/Sub — Derived Rules
 > Derives from `domains/messaging/first-principles.md`. Applies P1
 > (Messages are Contracts), P2 (Ordering is a Property, Not an
 > Assumption), P3 (Consumers are Idempotent), and P4 (Delivery
 > Semantics are Explicit) primarily, with P7 (partitioning), P10
 > (per-subscription lag). The `messaging-shared-subscription` chaos
 > anti-pattern lives here (pre-specified in P4 ATELIER-110).
 > Cross-links `domains/messaging/streams` for the pub/sub-vs-stream
 > durability boundary and `domains/observability/metrics` for
 > per-subscription lag.
 ## What Pub/Sub Is (P1 Messages are Contracts)
 - Pub/sub is the fan-out primitive: a producer publishes a message
  to a topic; N independent subscriptions each receive a copy. The
  message has an explicit, versioned schema (P1): the topic's
  schema is the contract every subscription agrees to before
  subscribing. A schemaless topic is a defect — every subscriber
  breaks silently on the next shape change.
 - The boundary with queues is the fan-out ratio. A queue is
  point-to-point (one producer, one consumer); pub/sub is
  one-to-many (one producer, N consumers, each with its own
  subscription). The boundary with streams is the durability model
  — see the cross-link below. Pub/sub is an async concern because
  producer and consumers are separate systems and the failure model
  is network, not crash (D-062).
 - See `domains/messaging/queues.md` for the point-to-point variant
  and `domains/messaging/streams.md` for the durable-log variant.
 ## Topic / Subscription Model (P1, P3, P4)
 - A **topic** is the named stream of messages. A **subscription**
  is a durable cursor over the topic: each subscription receives
  every message published after it was created (subject to
  retention and filtering). The subscription is independent — its
  ack, redelivery, and DLQ are per-subscription, not shared.
 - Each subscription is a consumer under at-least-once by default
  (P4): the broker redelivers until the subscription acks, and the
  subscriber must be idempotent (P3). A subscription with no
  idempotency dedup duplicates every redelivered message.
 - The topic's schema evolves compatibly (P9 — Schemas Evolve
  Compatibly): a new field the old subscriber ignores is
  backward-compatible; a renamed field the old subscriber parses
  as `undefined` is a P1 violation.
 ```python
 # Publish + two independent subscriptions (P1 contract, P3
 # idempotency, P4 at-least-once per subscription). Each
 # subscription is an independent durable cursor; acking one does
 # not affect the other.
 import json, uuid
 # --- Publisher ---
 def publish(topic, event, broker):
    # P1: versioned schema on the topic. All subscribers must
    # understand this schema (or a compatible superset — P9).
    message = {
        "schema": "user.signed-up.v1",
        "id": str(uuid.uuid4()),
        "idempotencyKey": f"user:{event['userId']}:signup",
        "payload": event,
    }
    broker.publish(topic=topic, body=json.dumps(message))
 # --- Subscription A: welcome-email service ---
 def subscribe_welcome(broker, dedup_store, send_email):
    sub = broker.subscribe(topic="users", subscription="welcome-email")
    for message in sub.receive():
        # P3: idempotent per subscription. A redelivered message is
        # a no-op for THIS subscription, not for the others.
        if dedup_store.seen(("welcome", message["idempotencyKey"])):
            sub.ack(message)
            continue
        send_email(message["payload"]["email"], "Welcome!")
        dedup_store.mark(("welcome", message["idempotencyKey"]))
        sub.ack(message)
 # --- Subscription B: analytics-ingest service ---
 def subscribe_analytics(broker, dedup_store, ingest):
    # Independent subscription: its own cursor, its own dedup,
    # its own ack. Welcome-email acking does NOT advance this.
    sub = broker.subscribe(topic="users", subscription="analytics")
    for message in sub.receive():
        if dedup_store.seen(("analytics", message["idempotencyKey"])):
            sub.ack(message)
            continue
        ingest(message["payload"])
        dedup_store.mark(("analytics", message["idempotencyKey"]))
        sub.ack(message)
 ```
 - The dedup key is scoped per subscription: `(subscription,
  idempotencyKey)`. A redelivery to subscription A that was already
  processed by A is a no-op for A; the same message delivered to
  subscription B is processed by B independently. Scoping the dedup
  key by subscription prevents one subscription's dedup from
  masking another's redelivery.
 ## Fan-Out Semantics (P4, P7)
 - Fan-out means every subscription receives every published message
  (subject to filtering — see below). The broker duplicates the
  message per subscription; each subscription's delivery is
  independent. The fan-out ratio is the number of subscriptions; the
  broker's cost scales with fan-out × message size.
 - Partitioning (P7) applies to topics that are partitioned for
  throughput: a partitioned topic delivers per-partition order, and
  each subscription receives from every partition. A subscription
  that consumes partitions in parallel must handle per-partition
  ordering and cross-partition non-ordering (P2 — document the
  property, do not assume global order).
 - The delivery semantic is per-subscription (P4): subscription A
  may be at-least-once, subscription B may be at-most-once (for a
  loss-tolerant analytics feed). The choice is per subscription,
  declared, not emergent.
 ## Shared vs Independent Subscriptions (P2, P3 — the chaos anti-pattern)
 - An **independent subscription** is one durable cursor per
  consumer group: each subscription receives every message in
  topic order (per partition, P2) and acks independently. This is
  the correct default: per-consumer ordering and per-consumer
  idempotency hold.
 - A **shared subscription** is one subscription shared by multiple
  consumers: the broker dispatches each message to an arbitrary
  consumer in the shared group. This breaks per-consumer ordering
  (P2 — consumer A sees message 3 before consumer B sees message
  1) and complicates idempotency (P3 — the dedup state must be
  shared across consumers, not per-consumer). This is the
  `messaging-shared-subscription` chaos anti-pattern
  (pre-specified in P4 ATELIER-110): the primary breach is P2
  (ordering); P3 (idempotency) is the compounding consequence.
 - A shared subscription is correct ONLY when the consumers are
  stateless, the per-message processing is order-independent, and
  the dedup store is shared (a shared Redis, a shared DB). A shared
  subscription for order-dependent or per-consumer-stateful
  processing is the chaos anti-pattern: the broker's arbitrary
  dispatch breaks the order the consumer assumes.
 ```python
 # The messaging-shared-subscription chaos anti-pattern (P2
 # ordering breach, P3 idempotency compounding). Two consumers
 # share one subscription; the broker dispatches each message to
 # an arbitrary consumer. Per-consumer ordering breaks; dedup
 # must be shared (and often is not).
 # BAD — shared subscription, per-consumer dedup (chaos):
 def shared_subscription_bad(broker, send_email):
    # Both consumers call subscribe with the SAME subscription
    # name. The broker round-robins; consumer A gets msg 1, msg 3;
    # consumer B gets msg 2, msg 4. Per-consumer order is broken.
    # If each consumer has its OWN dedup store, a redelivery to
    # the OTHER consumer re-processes (P3 breach).
    sub = broker.subscribe(topic="users", subscription="shared")
    for message in sub.receive():
        # Per-consumer dedup — WRONG. A redelivered message may
        # land on the other consumer, which has not seen it.
        if local_dedup.seen(message["idempotencyKey"]):  # per-consumer
            sub.ack(message); continue
        send_email(message["payload"]["email"], "Welcome!")
        local_dedup.mark(message["idempotencyKey"])
        sub.ack(message)
 # CORRECT — independent subscriptions (per-consumer ordering,
 # per-subscription dedup):
 def independent_subscriptions_good(broker, send_email):
    sub = broker.subscribe(topic="users", subscription="welcome-email")
    for message in sub.receive():
        if dedup_store.seen(("welcome", message["idempotencyKey"])):
            sub.ack(message); continue
        send_email(message["payload"]["email"], "Welcome!")
        dedup_store.mark(("welcome", message["idempotencyKey"]))
        sub.ack(message)
 ```
 - If a shared subscription is genuinely required (stateless,
  order-independent, shared dedup), document the choice and the
  shared-dedup requirement (P2 — the ordering property is "none
  across consumers"; P3 — the dedup is shared). The default is
  independent subscriptions; shared is an opt-in for the narrow case.
 ## Filtering (P4, C8 Economy)
 - **Subscription filtering** lets a subscription receive only
  messages matching a filter (e.g., `event.type == "order"`).
  Filtering at the broker saves bandwidth (C8 — the subscriber
  does not receive and discard) and reduces subscriber load.
 - **Server-side filtering** (broker evaluates the filter before
  delivery) is more efficient than **client-side filtering**
  (subscriber receives and discards). Server-side filtering is the
  default where the broker supports it (GCP Pub/Sub, SNS filtering,
  NATS subject filtering); client-side is the fallback.
 - A filter that is too broad wastes bandwidth; a filter that is
  too narrow drops messages the subscriber needed. The filter is
  a P1 (contract) and P4 (semantic) decision: the subscription's
  filter is part of its declared contract.
 ## Ordering Across Subscriptions (P2)
 - A topic with per-partition ordering delivers per-partition order
  to each subscription. Across subscriptions, there is no ordering
  guarantee: subscription A may ack message 3 while subscription B
  is still on message 1. This is correct and expected — each
  subscription is independent.
 - Within a subscription, ordering holds per partition (P2 — the
  documented property). A subscription that processes partitions
  in parallel must not assume cross-partition order. A subscription
  that needs global order must use a single partition (sacrificing
  parallelism, P7) or an external sequencing mechanism.
 - The `messaging-shared-subscription` anti-pattern breaks even
  per-partition order within a subscription: the broker's arbitrary
  dispatch to consumers in the shared group breaks the per-
  partition sequence each consumer sees.
 ## Pub/Sub vs Stream — The Durability Boundary (cross-link messaging/streams)
 - Pub/sub and streams are both fan-out or one-to-many primitives,
  but their durability model differs. Pub/sub is a
  **push-to-subscription** model: each subscription is a cursor,
  retention is short (the subscription's unacked window), and
  replay is limited to the unacked messages. A subscription that
  falls behind beyond the retention window loses messages
  permanently.
 - A stream is a **durable-log** model: messages are retained by
  the log for a configured window (P8 — Replay and Retention are
  Configured), and any consumer group can replay from any offset
  within the window. A stream consumer that falls behind can
  catch up by replaying; a pub/sub subscription that falls behind
  beyond retention cannot.
 - The choice is the durability requirement: if the consumer must
  be able to replay (reprocessing, backfill, new consumer starting
  from the beginning), use a stream. If the consumer only needs
  the live feed (and can tolerate loss on a long fall-behind),
  pub/sub is lighter. See `domains/messaging/streams.md` for the
  durable-log model, offsets, and consumer groups.
 ## Observability — Per-Subscription Lag (P10)
 - Per-subscription lag (messages published minus messages acked
  for each subscription, or the age of the oldest unacked message
  per subscription) is the primary pub/sub health signal. Each
  subscription has its own lag — a fast subscription and a slow
  subscription on the same topic are independent signals.
 - A subscription whose lag grows beyond the retention window is a
  silent data-loss risk: the broker will drop the oldest messages,
  and the subscription will never see them. Alert on lag relative
  to retention — lag approaching retention is the loss threshold.
 - Wire per-subscription lag to `domains/observability/metrics` as
  an SLI per subscription. A topic with N subscriptions has N lag
  metrics; a single aggregate hides the slow one. See
  `domains/observability/metrics` for the generic SLI/SLO
  discipline.
 ## What Violates Pub/Sub Discipline
 | Violation | Principle |
 |-----------|-----------|
 | Shared subscription for order-dependent processing (broker dispatch breaks per-consumer order) | P2 Ordering is a Property, Not an Assumption |
 | Shared subscription with per-consumer dedup (redelivery to the other consumer re-processes) | P3 Consumers are Idempotent |
 | Schemaless topic (no versioned contract; subscribers parse by guess) | P1 Messages are Contracts |
 | Subscription with no idempotency dedup (redelivered message duplicates the effect) | P3 Consumers are Idempotent |
 | Subscription whose lag approaches retention (silent data loss) | P10, `domains/observability/metrics` |
 | Unstated delivery semantic per subscription (at-least-once vs at-most-once guessed) | P4 Delivery Semantics are Explicit |
 | Filter that is too narrow (drops messages the subscriber needed) | P1, P4 |
 | Partitioned topic with no documented per-partition ordering contract | P2 Ordering is a Property, Not an Assumption |
 | No per-subscription lag metric (slow subscription invisible) | P10 Messaging is Observable |
 | Cross-partition order assumption within a subscription (no global order guarantee) | P2, P7 |
@@ -0,0 +1,318 @@
 # Queues — Derived Rules
 > Derives from `domains/messaging/first-principles.md`. Applies P1
 > (Messages are Contracts), P3 (Consumers are Idempotent), P4
 > (Delivery Semantics are Explicit), P5 (Dead-Letter Handling is
 > Defined), and P6 (Backpressure is Bounded) primarily, with P2
 > (ordering), P7 (partitioning), and P10 (observable lag). For the
 > at-least-once / at-most-once / exactly-once decision, see the
 > comparison table below. Cross-links `domains/concurrency/patterns`
 > for the in-process bounded-queue analog and
 > `domains/observability/metrics` for consumer lag.
 ## What a Queue Is (P1 Messages are Contracts)
 - A queue is a point-to-point async delivery primitive. A producer
  enqueues a message; exactly one consumer dequeues and processes
  it. The message has an explicit, versioned schema (P1): the
  producer and consumer agree on shape before exchange, and the
  schema is the boundary — a schemaless message is a defect (the
  consumer breaks silently on the next shape change).
 - The boundary is per D-062: messaging owns the cross-process /
  network-failure-model angle; concurrency owns the in-process
  analog. A queue is a messaging concern because producer and
  consumer are separate systems, the broker is the intermediary,
  and the failure model is network (the message can be lost,
  duplicated, reordered, or delayed by the broker or the network,
  not by a thread crash). The in-process bounded buffer
  (`domains/concurrency/patterns` Pattern 5) is the analog below
  this boundary — it fails by OOM; a broker-backed queue fails by
  partition, broker restart, or consumer crash-and-retry.
 - See `domains/messaging/pubsub.md` for the fan-out (one-to-many)
  variant and `domains/messaging/streams.md` for the durable-log
  (replay-from-offset) variant. A queue deletes on ack; a stream
  retains for replay — the durability boundary is the
  distinguishing trait.
 ## Producer / Consumer Model (P1, P3)
 - The producer enqueues a message with an idempotency key (P3).
  The consumer dequeues, processes, and acks. If the consumer
  crashes before acking, the broker redelivers; the idempotency
  key makes the redelivery safe (the consumer dedups, not the
  broker).
 - The idempotency key is per-message, not per-producer or
  per-session. A consumer that dedups by producer alone will drop
  distinct messages issued in the same window. Use a UUID per
  message, or a deterministic key derived from the message content
  (e.g., `(entity, operation, version)`).
 ```python
 # Producer/consumer pair with idempotency key (P1 contract, P3
 # idempotency). The producer tags each message with a versioned
 # schema and a unique idempotency key; the consumer dedups by the
 # key so a redelivered message is processed once (P3).
 # --- Producer ---
 import json, uuid
 def enqueue(order, broker):
    # P1: versioned schema. The message carries its schema version
    # so the consumer can route by shape (P9 evolution discipline).
    message = {
        "schema": "order.created.v1",
        "id": str(uuid.uuid4()),
        "idempotencyKey": f"order:{order['id']}:{order['version']}",
        "payload": order,
    }
    broker.send(queue="orders", body=json.dumps(message))
    # At-least-once by default (P4): the broker acks the send; the
    # consumer may see this message more than once under retry.
 # --- Consumer ---
 def consume(broker, dedup_store, process_order):
    for message in broker.receive(queue="orders"):
        # P3: idempotent consumer. Dedup by idempotency key before
        # processing; a redelivered message is a no-op, not a
        # double-apply.
        if dedup_store.seen(message["idempotencyKey"]):
            broker.ack(message)        # already processed; skip
            continue
        try:
            process_order(message["payload"])
            dedup_store.mark(message["idempotencyKey"])
            broker.ack(message)        # success; broker drops it
        except Exception:
            broker.nack(message)       # redeliver (at-least-once, P4)
 ```
 - The dedup store is bounded (P6 — Backpressure is Bounded): a
  dedup store that grows without bound is a memory leak. Use a TTL
  window longer than the broker's max-redelivery window, or a
  bounded LRU. See `domains/messaging/delivery-semantics.md` for
  the full idempotency-key dedup-store pattern.
 ## Ack / Nack (P4 Delivery Semantics are Explicit)
 - **Ack** tells the broker the message was processed; the broker
  drops it. **Nack** (negative ack) tells the broker the
  processing failed; the broker redelivers (at-least-once) or
  routes to a DLQ (after the retry budget — P5).
 - A consumer that neither acks nor nacks within the visibility
  timeout causes the broker to redeliver (the broker assumes the
  consumer died). This is the at-least-once default: the broker
  prefers duplication to loss.
 - The semantic is explicit (P4): at-least-once is the default; the
  consumer must be idempotent (P3). At-most-once is fire-and-forget
  (no ack; the broker drops on send) — lossy but lowest latency.
  Exactly-once is at-least-once plus idempotency, or a
  transactional two-phase commit — see the comparison table below.
 ## Visibility Timeouts and Redelivery (P4, P5)
 - The visibility timeout is the window the broker hides a message
  after delivery, waiting for the ack. If the consumer does not
  ack within the window, the broker makes the message visible
  again and redelivers it (to the same consumer or another). This
  is the at-least-once mechanism: the broker assumes a
  no-ack-in-time consumer is dead.
 - The timeout must be longer than the processing time, or the
  broker redelivers a message the consumer is still processing —
  causing duplicate processing (which P3 idempotency makes safe,
  but which wastes resources). A timeout shorter than processing
  time is a P6 (backpressure) smell: the consumer is too slow for
  the configured timeout.
 - Redelivery has a budget (P5): after N redeliveries or a TTL, the
  message routes to the DLQ. An unbounded retry budget is the
  `messaging-unbounded-retry` chaos anti-pattern: the consumer
  never makes progress past the poison message.
 ## FIFO vs Standard Queues (P2 Ordering is a Property, Not an Assumption)
 - A **standard queue** delivers in arrival order per receive-node
  but offers no global ordering across shards, no per-message-
  group ordering, and may redeliver out of order under retry. It
  is the high-throughput default; ordering is *not* guaranteed
  (P2: the ordering property is "none" — explicitly documented).
 - A **FIFO queue** delivers strict per-message-group order: all
  messages with the same group ID are delivered to one consumer
  in send order. The cost is throughput (FIFO queues cap at lower
  TPS) and latency (the broker must sequence per group). The
  ordering property is "per-group strict" — explicitly documented
  (P2).
 - The choice is a P2 decision (which ordering guarantee) and a C8
  decision (throughput cost). A consumer that assumes FIFO on a
  standard queue is a P2 violation: the broker does not provide
  the guarantee the consumer assumes. Document the property; do
  not assume it.
 ## Prefetch and Concurrency (P6 Backpressure is Bounded)
 - **Prefetch** (or max-unacked) bounds how many messages the
  broker delivers to one consumer without an ack. A prefetch of 1
  is strict stop-and-wait (lowest throughput, tightest backpressure);
  a prefetch of N allows the consumer to process N in flight
  (higher throughput, more memory). An unbounded prefetch is a P6
  violation: the broker floods the consumer's memory.
 - **Consumer concurrency** is the number of parallel workers
  processing from the queue. More workers increase throughput up to
  the downstream's limit; beyond that, the workers stall the
  downstream (P6 — the backpressure propagates to the
  downstream, not the broker).
 - The prefetch × concurrency product is the in-flight cap. Declare
  it (P6): an undeclared cap is a defect — the consumer either
  underutilizes the broker (prefetch too low) or OOMs under load
  (prefetch too high). This is the cross-process analog of
  `domains/concurrency/patterns` Pattern 5 (bounded queue with
  backpressure): concurrency owns the in-process analog;
  messaging owns the broker-backed instance.
 ## Long Polling (P6, C8 Economy)
 - Long polling (or `ReceiveMessage` with a wait-time-seconds)
  holds the receive request open until a message arrives or the
  wait expires. This reduces empty-receive round trips (C8
  economy of the constrained link) and reduces latency-to-first-
  message (the message is delivered when it arrives, not on the
  next poll cycle).
 - Long polling is the default for low-throughput queues: short
  polling burns CPU on empty receives; long polling waits for
  work. For high-throughput queues, the broker is usually full
  enough that long polling adds no latency; for low-throughput
  queues, long polling is the difference between 20ms and 20s
  latency-to-first-message.
 ## Redelivery + DLQ Flow (P5 Dead-Letter Handling is Defined)
 - A poison message (unparseable, repeatedly failing, or exhausting
  the retry budget) routes to the dead-letter queue. The DLQ is
  observable (P10 — DLQ depth is an alert) and drainable (an
  operator can inspect, replay, or discard with audit).
 - The retry budget is bounded (P6): N redeliveries, or a TTL with
  exponential backoff. After the budget is exhausted, the message
  is moved to the DLQ, not retried forever. An unbounded retry is
  the `messaging-unbounded-retry` chaos anti-pattern (P5 breach).
 - The DLQ routing rule is a redelivery-count or TTL threshold
  plus a target queue. See `domains/messaging/delivery-semantics.md`
  for the dead-letter strategy comparison table.
 ```python
 # Redelivery + DLQ flow (P5 dead-letter handling, P6 bounded
 # retry budget). The consumer tracks redelivery count; after the
 # budget, the message routes to the DLQ. The DLQ is observable
 # (P10 — depth is an alert) and drainable.
 MAX_REDELIVERIES = 5
 DLQ = "orders-dlq"
 def consume_with_dlq(broker, dedup_store, process_order):
    for message in broker.receive(queue="orders"):
        # P3 idempotency: a redelivered, already-processed message
        # is acked and skipped (not re-processed, not DLQ'd).
        if dedup_store.seen(message["idempotencyKey"]):
            broker.ack(message)
            continue
        try:
            process_order(message["payload"])
            dedup_store.mark(message["idempotencyKey"])
            broker.ack(message)
        except Exception as exc:
            # P5: bounded retry budget. After MAX_REDELIVERIES,
            # route to DLQ — do NOT retry forever.
            count = message.get("redeliveryCount", 0) + 1
            if count >= MAX_REDELIVERIES:
                broker.send(DLQ, body=json.dumps({
                    "original": message,
                    "reason": str(exc),
                    "deadLetteredAt": now_iso(),
                    "redeliveryCount": count,
                }))
                broker.ack(message)  # remove from the origin queue
                # P10: the DLQ depth must alert. A DLQ that grows
                # with no alert is a silent correctness defect.
            else:
                # Nack with backoff: the broker redelivers after a
                # delay. The backoff caps the retry rate (P6).
                broker.nack(message, delay=exponential_backoff(count))
 ```
 - The `deadLetteredAt` and `reason` fields make the DLQ entry
  observable and auditable: an operator inspecting the DLQ sees
  why each message was dead-lettered and when. See
  `domains/errors/patterns` for the errors-as-data discipline the
  DLQ entry follows.
 ## Delivery Semantics Comparison (D-069)
 | Semantic | Guarantee | Latency Cost | Implementation Cost | When It Fits |
 |----------|-----------|-------------|---------------------|--------------|
 | **At-most-once** | A message is delivered 0 or 1 times; loss is possible, duplication is not | Lowest (no ack; fire-and-forget) | Lowest (no ack, no dedup) | Telemetry where a dropped sample is acceptable; high-throughput metrics; MQTT QoS 0; logs where a lost line is tolerable. Never for billing, orders, or any side-effecting operation. |
 | **At-least-once** | A message is delivered 1 or more times; duplication is possible, loss is not | Low (one ack round-trip) | Medium (consumer must be idempotent — P3; dedup store required) | The default for side-effecting operations: orders, payments, commands. The consumer dedups via idempotency keys (P3); the broker guarantees delivery. Fits the vast majority of broker-backed queues (SQS standard, RabbitMQ ack, MQTT QoS 1). |
 | **Exactly-once** | A message is delivered exactly 1 time; no loss, no duplication | Highest (two-phase commit or transactional producer+consumer) | Highest (requires transactions, a transactional producer, and a transactional consumer — or at-least-once plus idempotency, which collapses to at-least-once with dedup) | Rare. Kafka transactions (consume-process-produce in one transaction); MQTT QoS 2 (four-step handshake). In practice, "exactly-once" is usually at-least-once plus idempotency (P3) — the broker does not guarantee it; the consumer enforces it. Jepsen analyses verify broker claims. |
 - The default for side-effecting operations is **at-least-once with
  idempotent consumers** (P3). At-most-once is for loss-tolerant
  telemetry. Exactly-once is reserved for the narrow case where
  the consume-process-produce loop must be transactional (Kafka
  transactions) — and even then, the consumer should be idempotent
  as defense-in-depth.
 - The latency cost column is the C8 tradeoff: at-most-once is
  cheapest, exactly-once is most expensive. The implementation
  cost column is the C1/C3 tradeoff: at-most-once is simplest,
  exactly-once is most complex (and most fragile — a transactional
  consumer that partially fails is a bug source). The "when it
  fits" column is the P4 decision: declare the semantic per
  channel, do not let it emerge.
 - This is the decision matrix required by D-069 for queues; see
  `domains/messaging/delivery-semantics.md` for the correctness
  properties of each semantic and the dead-letter strategy
  comparison.
 ## Observability — Consumer Lag (P10 Messaging is Observable)
 - Consumer lag (messages enqueued minus messages acked, or the
  age of the oldest unacked message) is the primary queue health
  signal. A lag that grows without bound is a P6 violation (the
  consumer is slower than the producer) and a P10 violation if it
  is not alerted.
 - DLQ depth is the secondary signal: a DLQ that grows is a P5
  signal (poison messages are accumulating) and a P10 signal if
  not alerted. Wire both to `domains/observability/metrics` as
  SLIs with SLOs (e.g., lag < 1000 messages, DLQ depth < 10).
 - A queue with no lag metric is operating blind (P10 violation):
  the operator cannot see the consumer falling behind until the
  downstream effect surfaces.
 ## Cross-Link to Concurrency (P6, cross-link concurrency/patterns)
 - The broker-backed queue is the cross-process analog of the
  in-process bounded buffer. `domains/concurrency/patterns`
  Pattern 5 (Bounded Queue with Backpressure) owns the in-process
  instance ("producer is blocked or signaled" within one program);
  messaging owns the broker-backed instance above it (the producer
  is the broker's enqueue, the consumer is the broker's dequeue, the
  backpressure is the prefetch cap and the lag signal). The
  failure model differs: in-process fails by OOM; broker-backed
  fails by network partition, broker restart, or consumer
  crash-and-retry (D-062).
 - The cross-link is one-directional outward (messaging →
  concurrency) per D-026 extended: messaging references concurrency
  as the in-process foundation; concurrency does not back-link to
  messaging.
 ## What Violates Queue Discipline
 | Violation | Principle |
 |-----------|-----------|
 | Schemaless message (no versioned contract; consumer parses by guess) | P1 Messages are Contracts |
 | Non-idempotent consumer under at-least-once delivery (redelivery doubles the effect) | P3 Consumers are Idempotent |
 | Unstated delivery semantic (at-least-once vs exactly-once guessed) | P4 Delivery Semantics are Explicit |
 | No DLQ (poison message retried forever or silently dropped) | P5 Dead-Letter Handling is Defined |
 | Unbounded prefetch (broker floods consumer memory) | P6 Backpressure is Bounded |
 | Unbounded retry budget (no cap; consumer never progresses past the poison) | P5, P6 |
 | Consumer that assumes FIFO on a standard queue (ordering not guaranteed) | P2 Ordering is a Property, Not an Assumption |
 | Visibility timeout shorter than processing time (redeliver while still processing) | P4, P6 |
 | DLQ with no depth alert (poison messages accumulate invisibly) | P10, `domains/observability/metrics` |
 | No consumer-lag metric (consumer falls behind invisibly) | P10, `domains/observability/metrics` |
 | Dedup store that grows without bound (memory leak) | P6 Backpressure is Bounded |
 | Default prefetch with no rationale (underutilizes or OOMs) | P6, `domains/concurrency/patterns` |
@@ -0,0 +1,329 @@
 # Streams — Derived Rules
 > Derives from `domains/messaging/first-principles.md`. Applies P8
 > (Replay and Retention are Configured) primarily, with P2
 > (per-partition ordering), P3 (idempotent consumers), P4 (exactly-
 > once via transactions), P7 (partitioning), and P10 (observable
 > consumer-group health). For the Kafka / Kinesis / Pulsar /
 > NATS JetStream decision, see the stream-platform comparison
 > table below. For the consumer-group rebalance strategy choice,
 > see the rebalance enumeration below (IDEATE-41). Cross-links
 > `domains/messaging/delivery-semantics` for exactly-once via
 > transactions, `domains/data/schema-design` for stream schema,
 > and `domains/observability/tracing` for cross-partition traces.
 ## What a Stream Is (P8 Replay and Retention are Configured)
 - A stream is a durable-log messaging primitive. Messages are
  appended to a partitioned, replicated log; consumers read from an
  offset and advance at their own pace. The log is retained for a
  configured window (P8) — a stream is a durable log, not a pipe.
  A consumer that falls behind can catch up by replaying from an
  earlier offset; a consumer that starts fresh can replay from the
  beginning (within retention).
 - The boundary with queues and pub/sub is the durability model. A
  queue deletes on ack; a pub/sub subscription retains only its
  unacked window; a stream retains the whole log for the configured
  retention. This makes a stream replayable (P8 — the
  reversibility mechanism) and observable as a record (P10 — the
  log is itself an audit of what happened). See
  `domains/messaging/pubsub.md` for the pub/sub-vs-stream
  durability boundary discussion.
 - The boundary with concurrency is per D-062: messaging owns the
  cross-process/network-failure-model angle. A stream is a
  messaging concern because the log spans brokers and consumers
  across a network, and the failure model is partition, broker
  restart, or consumer crash-and-retry — not in-process OOM or
  thread crash.
 ## Partitioned Log Model (P2, P7)
 - A stream is partitioned for throughput and parallelism. Each
  partition is an ordered, append-only log; messages within a
  partition are strictly ordered (P2 — per-partition strict order
  is the documented property). Across partitions, there is no
  ordering guarantee: partition 0 and partition 1 are independent
  logs.
 - The partition key (P7 — Partitioning is Intentional) determines
  which partition a message lands on. A key that hashes evenly
  spreads load; a key that matches the per-entity ordering need
  (e.g., `userId` for user events) keeps a user's events on one
  partition in order; a key that is too coarse (one partition for
  the whole topic) serializes all traffic.
 - The partition count is a capacity bound: it caps the parallelism
  (one consumer per partition per consumer group) and the
  throughput (each partition has a write-throughput limit). Too
  few partitions cap parallelism; too many partition overhead the
  broker (file handles, replication, rebalance cost). The choice
  is documented (P7), not defaulted.
 ## Offsets (P2, P8)
 - An offset is a consumer's position in a partition. The consumer
  reads from its last committed offset; acking (committing the
  offset) advances it. A consumer that crashes before committing
  re-reads from the last committed offset (at-least-once by
  default, P4) — the consumer must be idempotent (P3).
 - The offset is per-partition (P2): each partition has its own
  position, and the consumer commits them independently (or
  atomically across partitions in a transaction — see below).
 - Replay (P8) is resetting the offset backward: a consumer can
  replay from the beginning of retention, from a timestamp, or
  from a specific offset. This is the durable-log property that
  distinguishes a stream from a queue.
 ## Consumer Groups (P3, P7, P10)
 - A consumer group is a set of consumers sharing the stream's
  partitions: each partition is assigned to exactly one consumer
  in the group. The group is the unit of parallelism and the unit
  of offset tracking. Within a group, each consumer handles its
  assigned partitions; across groups, each group independently
  reads the whole stream (the pub/sub fan-out property, per
  subscription/group).
 - A consumer in the group is idempotent (P3): under at-least-once
  (the default), a redelivery after a crash-and-retry re-processes
  messages. The consumer dedups by idempotency key, or processes
  deterministically (e.g., a stateful aggregation that overwrites
  with the latest value).
 - The consumer group's health is observable (P10): per-partition
  lag (offset of the consumer vs the log's head), the group's
  consumption rate, and rebalance events are first-class signals.
  A group whose lag grows without bound is a P6 (backpressure)
  smell and a P10 (observability) violation if not alerted.
 ```python
 # Consumer-group reading from offsets (P2 per-partition order,
 # P3 idempotent under at-least-once, P7 partition assignment,
 # P8 replay from offset). Each consumer in the group handles its
 # assigned partitions; the group commits offsets atomically or
 # per-partition.
 def consume_stream(stream, group, dedup_store, process_event):
    # Assign partitions to this consumer by the group's
    # rebalance strategy (see the enumeration below).
    for partition in stream.assigned_partitions(group, consumer=ME):
        # Read from the last committed offset (P8 — replay by
        # resetting this offset).
        offset = stream.committed_offset(group, partition)
        for message in stream.read(partition, from_offset=offset):
            # P3: idempotent under at-least-once. A redelivery
            # after a crash-and-retry re-processes; dedup by key.
            if dedup_store.seen(message["idempotencyKey"]):
                stream.commit(group, partition, message["offset"])
                continue
            process_event(message["payload"])
            dedup_store.mark(message["idempotencyKey"])
            # Commit the offset to advance (P8 — the position is
            # the replay pointer).
            stream.commit(group, partition, message["offset"])
 ```
 - The commit-after-process order gives at-least-once (a crash
  before commit re-reads); the commit-before-process order gives
  at-most-once (a crash after commit loses the unprocessed
  message). The default is at-least-once with idempotent consumers
  (P3, P4).
 ## Consumer-Group Rebalance Strategies (IDEATE-41, ATELIER-100 refinement)
 When a consumer joins or leaves the group, the broker must
 reassign partitions. The rebalance strategy determines the cost
 and the use-case fit. This enumeration parallels the v0.3
 IDEATE-30 drift-type enumeration (each strategy with its
 stop-the-world cost and use-case fit).
 | Strategy | Mechanism | Partition Stop-the-World Cost | Use-Case Fit |
 |----------|-----------|-------------------------------|--------------|
 | **Eager rebalance** (stop-the-world) | Every consumer in the group revokes ALL its partitions, the broker reassigns the full partition set, then consumers resume. Every rebalance pauses the whole group. | High — every partition pauses for every rebalance; the whole group stops processing during the revocation+reassignment window. Throughput drops to zero during rebalance. | Simple brokers, small groups, or rarely-rebalancing groups where the simplicity of full revocation outweighs the pause cost. Kafka's legacy protocol (pre-2.4). Avoid for large groups or frequent scale events. |
 | **Sticky (incremental cooperative) rebalance** | The broker reassigns only the partitions that must move (the joining/leaving consumer's share); existing partitions stay assigned. The rebalance is incremental and cooperative — no full revocation. | Low — only the moving partitions pause; the rest of the group continues processing. The pause is proportional to the changed partition count, not the total. | The default for large groups, frequent scale events, and rolling deploys. Kafka's CooperativeStickyAssignor (2.4+), Pulsar, NATS JetStream. Prefer for any group where a full stop-the-world on every deploy is unacceptable. |
 | **Cooperative (no-revoke) rebalance** | A subset of sticky where no partition is revoked unless the consumer leaves; only additions are incremental. The strictest minimization of stop-the-world. | Lowest — only added partitions pause; existing assignments are untouched. | Groups where partition assignment is append-only (consumers join but rarely leave). Useful for long-lived consumers with incremental scaling. |
 - The default for any non-trivial group is **sticky/cooperative**:
  a rolling deploy that triggers an eager rebalance pauses the
  whole group on every pod restart, which is unacceptable at
  scale. The eager strategy is a legacy default that survives
  because it is simple; prefer sticky where the broker supports
  it.
 - The "partition stop-the-world cost" column is the P6
  (backpressure) angle: a full stop-the-world during rebalance
  causes lag to spike (the consumer is paused, the producer is
  not). Sticky rebalance bounds the spike to the moving partitions.
 - A rebalance that pauses without a lag alert is a P10 violation:
  the operator cannot see the rebalance-induced lag. Wire rebalance
  events to `domains/observability/metrics` as an event signal.
 ## Replay and Retention Windows (P8, C5 Reversibility)
 - Retention is the configured window the log keeps messages: time-
  based (e.g., 7 days), size-based (e.g., 10 GB per partition), or
  compacted (keep the latest value per key — a changelog). A
  stream with no retention is a pipe, not a log (P8 violation); a
  stream with infinite retention grows until disk exhaustion (P6
  violation — backpressure on the broker).
 - Replay (P8) is resetting a consumer's offset to re-read from
  within the retention window. Use cases: reprocessing after a
  consumer bug (replay from the timestamp of the buggy deploy),
  backfilling a new consumer (replay from the beginning), or
  reindexing (replay to rebuild a derived store).
 - Compacted topics (Kafka log-compaction, Pulsar compaction) keep
  the latest value per key and discard older values for the same
  key. This turns the log into a changelog — a durable
  materialized view that replays to the current state. Compaction
  is a P8 (retention) and C5 (reversibility) mechanism: the log
  retains the current state per key and is replayable to it.
 ## Stream Processing (P2, P3, P4)
 - Stream processing is computing over the stream as it arrives:
  windowing (tumbling, sliding, session windows), joins (stream-
  stream, stream-table), aggregations (count, sum, per-key
  windows), and stateful transformations. The processing is
  per-partition ordered (P2 — a window over a partition is
  deterministic; a window across partitions is not unless the
  window is global).
 - Stream processing consumers are idempotent (P3): a redelivery
  after a crash re-processes a window; the aggregation must
  tolerate re-application (e.g., a sum is idempotent under replay
  if the window is keyed by offset range, not by wall time).
 - Exactly-once stream processing (P4) requires transactions: the
  consume-process-produce loop is one transaction — the input
  offset commit and the output produce are atomic. See the
  transactional exactly-once producer below.
 ## Exactly-Once via Transactions (P4, P3)
 - Exactly-once stream processing is at-least-once plus a
  transaction: the consumer commits the input offset and produces
  the output in one transactional operation. If the consumer
  crashes mid-transaction, neither the offset commit nor the
  output produce happens — the consumer re-reads from the last
  committed offset and re-processes (at-least-once), but the
  transaction ensures the output is produced exactly once.
 - This is NOT a broker guarantee of exactly-once delivery; it is
  at-least-once delivery plus idempotent/transactional processing
  (P3, P4). Jepsen analyses of Kafka transactions confirm the
  boundaries: the transaction is atomic within the broker, but
  the downstream sink must be transactional or idempotent too.
 - See `domains/messaging/delivery-semantics.md` for the full
  exactly-once-via-idempotency discussion.
 ```python
 # Transactional exactly-once producer (P4 exactly-once via
 # transactions, P3 idempotent produce). The consume-process-
 # produce loop is one transaction: the input offset commit and
 # the output produce are atomic. A crash mid-transaction rolls
 # both back; the consumer re-reads and re-processes.
 def consume_transform_produce(stream, group, txn_producer):
    # Begin a transaction. All produces and the offset commit in
    # this block are atomic (P4).
    with txn_producer.transaction() as txn:
        for partition in stream.assigned_partitions(group, ME):
            offset = stream.committed_offset(group, partition)
            for message in stream.read(partition, from_offset=offset):
                output = transform(message["payload"])
                # P3: idempotent produce. The txn producer
                # dedups by an epoch+sequence so a retried
                # transaction does not double-produce.
                txn.produce(
                    topic="enriched-events",
                    key=message["key"],
                    value=output,
                    idempotencyKey=message["idempotencyKey"],
                )
                # Commit the input offset within the same
                # transaction (P4 atomicity). A crash before
                # txn.commit() rolls this back; the consumer
                # re-reads from the prior offset.
                txn.commit_offset(group, partition, message["offset"])
        # txn.commit() makes the produces and the offset commit
        # visible atomically. A crash before this point aborts
        # both; a crash after is safe (idempotent produce — P3).
 ```
 - The transactional producer's idempotency (P3) is the defense
  against a retried transaction: the broker dedups the output by
  the producer's epoch and sequence so a re-commit does not
  double-produce. The transaction (P4) is the defense against a
  partial failure: the offset and the output commit together.
 ## Stream Schema (P1, P9, cross-link data/schema-design)
 - A stream's messages carry a versioned schema (P1). The schema
  evolves compatibly (P9): a new field the old consumer ignores is
  backward-compatible; a renamed field the old consumer parses as
  `undefined` is a P1 violation compounded as an evolution defect.
 - Stream schemas are often registered in a schema registry
  (Confluent, Apicurio) that enforces compatibility on produce.
  A producer that tries to publish an incompatible schema is
  rejected; the registry is the P1/P9 enforcement point.
 - See `domains/data/schema-design` for the generic schema-design
  discipline (Avro, Protobuf, JSON Schema); messaging owns the
  stream-specific instance — the registry, the per-topic
  compatibility mode, the consumer-side routing by schema
  version.
 ## Stream-Platform Comparison (D-069)
 | Axis | Apache Kafka | AWS Kinesis | Apache Pulsar | NATS JetStream |
 |------|--------------|-------------|---------------|----------------|
 | **Ordering** | Per-partition strict (P2); global only via single-partition topic | Per-shard strict; global only via single shard | Per-partition strict; global via single partition; also supports shared (out-of-order) subscriptions | Per-stream strict; per-subject ordering; global via single stream |
 | **Partitioning model** | Partitions (immutable count post-creation; increase requires recreate); key→partition by hash | Shards (reshardable: split/merge at runtime); key→shard by hash | Partitions (resizable; Pulsar's layered architecture separates compute from storage); key→partition by hash | Streams (subject-based; republish to resize); key→stream by subject |
 | **Replay / retention** | Time- or size-based retention; compaction (latest-per-key); replay from offset or timestamp | Time-based retention (24h–365d); replay from sequence number or timestamp; no compaction | Time- or size-based; compaction; replay from offset or timestamp; tiered storage (hot/warm/cold) | Time- or size-based; per-stream max-age; replay from sequence; no native compaction |
 | **Consumer groups** | Group-coordinated; offsets stored in an internal topic; eager (legacy) or sticky/cooperative (2.4+) rebalance | Enhanced fan-out consumers (per-shard HTTP/2 push); KCL for group coordination; no native group rebalance (shard is the unit) | Group-coordinated; shared or failover subscription modes; cooperative rebalance | Per-stream consumers; durable cursors; no native group rebalance (stream is the unit) |
 | **Exactly-once** | Transactions (KIP-98): consume-process-produce atomic; idempotent producer (KIP-516) | No native exactly-once; at-least-once with consumer-side dedup (P3) | Transactions: produce-ack atomic; idempotent producer | At-least-once by default; dedup window per stream (P3 idempotency) |
 | **Use-case fit** | High-throughput durable logs, stream processing (Kafka Streams, Flink), event sourcing, multi-consumer replay | AWS-native streaming, log ingestion, simple ETL within AWS; low operational burden | Cloud-native, geo-replication, tiered storage, mixed pub/sub + streaming; multi-tenant | Lightweight, low-latency, edge-friendly; NATS ecosystem; simpler ops than Kafka |
 | **Watch out for** | Partition count is fixed at creation (resize requires recreate + republish); rebalance cost on large groups; operational complexity | Shard limits per account; no compaction; retention cap at 365 days; AWS lock-in | Two-arch (BookKeeper + Brokers) operational complexity; smaller ecosystem | Smaller ecosystem; no native compaction; fewer stream-processing libraries |
 - The default for high-throughput durable logs with multi-consumer
  replay is **Kafka**; for AWS-native streaming, **Kinesis**; for
  geo-replicated multi-tenant or mixed pub/sub + streaming,
  **Pulsar**; for lightweight low-latency edge-friendly streaming,
  **NATS JetStream** (which also cross-links `domains/edge/iot`
  via MQTT parallels — see the edge↔messaging bidirectionality in
  `domains/messaging/first-principles.md` §4).
 - The ordering column is the P2 check: every platform provides
  per-partition/per-shard strict order; none provides global order
  across partitions except by single-partition. The replay/
  retention column is the P8 check: every platform retains for a
  configured window; replay is from offset or timestamp. The
  exactly-once column is the P4 check: Kafka and Pulsar provide
  transactions; Kinesis and JetStream rely on at-least-once plus
  consumer-side idempotency (P3).
 ## Cross-Partition Traces (P10, cross-link observability/tracing)
 - A stream-processing pipeline that fans out across partitions
  must propagate a trace context per event: the trace ID follows
  the event from source to processed output, even as the event
  crosses partition boundaries. Without cross-partion traces, a
  downstream error cannot be traced back to its source event.
 - See `domains/observability/tracing` for the generic distributed-
  tracing discipline (trace context propagation, span
  correlation). Messaging owns the stream-specific instance: the
  trace context is a message header, the span boundary is the
  consume-process-produce edge, and the cross-partition
  correlation is by trace ID (not by partition — partitions are
  independent logs, P2).
 - A stream processor with no trace propagation is a P10
  violation: the operator cannot trace a processed event back to
  its source. Wire the trace context into every produce and every
  consume.
 ## What Violates Stream Discipline
 | Violation | Principle |
 |-----------|-----------|
 | Stream with no retention (pipe, not log; no replay) | P8 Replay and Retention are Configured |
 | Infinite retention (grows until disk exhaustion) | P8, P6 |
 | Consumer group with eager rebalance at scale (full stop-the-world per deploy) | P6, IDEATE-41 (use sticky/cooperative) |
 | Non-idempotent stream consumer under at-least-once (redelivery re-processes the window) | P3 Consumers are Idempotent |
 | Default partition key (no rationale; hotspot or wrong-order) | P7 Partitioning is Intentional |
 | Partition count too low (caps parallelism) or too high (overhead) | P7 |
 | Cross-partition order assumption (no global order guarantee) | P2 Ordering is a Property, Not an Assumption |
 | Exactly-once claimed without transactional consume-process-produce (P4 violation) | P4 Delivery Semantics are Explicit |
 | Stream schema with no registry / no compatibility enforcement (silent shape break) | P1, P9, `domains/data/schema-design` |
 | No per-partition lag metric (consumer falls behind invisibly) | P10, `domains/observability/metrics` |
 | No cross-partition trace propagation (downstream error untraceable) | P10, `domains/observability/tracing` |
 | Compacted topic treated as a full log (old values already discarded) | P8, C1 (compaction is a retention mode, not a full log) |
@@ -0,0 +1,222 @@
 # Bad Example: Compliance Audit Log (Two Breaches)
 > An audit logging implementation that violates **two** Atelier
 > compliance principles in one example (per IDEATE-26, D-044):
 > **P1** (Audit Logs are Append-Only) — a mutable audit log with
 > routine `DELETE`/`UPDATE` "cleanup" — and **P9** (Secrets and
 > Sensitive Data are Redacted in Audit) — a database password leaked
 > into an audit record. Each violation is cited, then fixed.
 ## The Code
 ```python
 # audit_log.py — the audit sink, stored in a mutable Postgres table
 import psycopg2, datetime
 # P1 VIOLATION: the audit log is a regular mutable table. There is no
 # write-once protection, no immutable bucket, no hash-chaining.
 # Any DB user with UPDATE/DELETE can rewrite history.
 CREATE_TABLE = """
 CREATE TABLE audit_log (
    id           BIGSERIAL PRIMARY KEY,
    timestamp    TIMESTAMPTZ NOT NULL,
    event        TEXT NOT NULL,
    actor        TEXT NOT NULL,
    target       TEXT,
    payload      JSONB,
    request_id   TEXT
 );
 -- No row-level immutability. No trigger preventing UPDATE/DELETE.
 """
 def write_event(event, actor, target=None, payload=None, request_id=None):
    conn = psycopg2.connect(os.environ["DATABASE_URL"])
    conn.execute(
        "INSERT INTO audit_log (timestamp, event, actor, target, payload, request_id) "
        "VALUES (%s, %s, %s, %s, %s, %s)",
        (datetime.datetime.utcnow(), event, actor, target,
         json.dumps(payload), request_id),
    )
 # P1 VIOLATION (continued): "cleanup" that mutates the audit log.
 # A routine job deletes records older than 30 days to "save space"
 # and updates records to "fix typos in the actor field."
 def cleanup_audit_log():
    conn = psycopg2.connect(os.environ["DATABASE_URL"])
    # DELETE — an audit record is destroyed. This is tampering,
    # dressed as housekeeping.
    conn.execute("DELETE FROM audit_log WHERE timestamp < NOW() - INTERVAL '30 days'")
    # UPDATE — an audit record is rewritten. The "fix" is the
    # violation; the original actor is lost.
    conn.execute("UPDATE audit_log SET actor = 'admin' WHERE actor LIKE 'svc-%'")
 ```
 ```python
 # The call site that leaks a secret into the audit log.
 def read_config(key):
    # ... fetches a secret from the secrets manager ...
    value = secrets_manager.get(key)   # e.g. the raw DB password
    # P9 VIOLATION: the raw secret value is written into the audit
    # payload. The append-only log is now a secret store.
    write_event(
        event="config.read",
        actor="api-server",
        target={"kind": "secret", "id": key},
        payload={"value": value},            # <- the secret, in plaintext
        request_id=req.id,
    )
    return value
 ```
 The resulting audit record:
 ```json
 {
  "id": 48213,
  "timestamp": "2026-08-05T09:12:03Z",
  "event": "config.read",
  "actor": "api-server",
  "target": {"kind": "secret", "id": "db-password"},
  "payload": {"value": "p@ssw0rd-sup3r-s3cr3t-plaintext"},
  "request_id": "req_91c2"
 }
 ```
 A week later, the `cleanup_audit_log` job `DELETE`s this record (it
 is older than 30 days in the team's "retention" — which is actually a
 storage-economy decision, not a policy), and `UPDATE`s every
 `svc-*` actor to `admin`. The secret was in the log for a week,
 readable by anyone with `SELECT` on the table; now the record of it
 having been there is gone.
 ## What Makes It Bad
 ### Breach 1 — Mutable Audit Log (Compliance P1 Audit Logs are Append-Only)
 - The audit log is a regular mutable Postgres table. `DELETE FROM
  audit_log WHERE timestamp < ...` and `UPDATE audit_log SET actor =
  ...` both succeed. The log is a draft, not a record.
 - Routine `DELETE` as "cleanup" is the cardinal P1 violation: the
  deletion of an audit record is itself an auditable incident, not a
  housekeeping task. "We deleted old records to save space" is a P3
  (Retention is Policy, Not Storage) violation *and* a P1 violation —
  the retention decision is driven by storage cost, and the
  mechanism is tampering.
 - The `UPDATE` that rewrites `svc-deploy` → `admin` destroys
  attribution (a P7 violation stacked on the P1 violation): the
  original actor is lost, and the replacement (`admin`) is a shared
  identity that could be any of ten engineers.
 - **Fix:** the audit sink is append-only *by construction*, not by
  policy. Write-once storage (WORM bucket, immutable log stream,
  hash-chained ledger) enforces immutability at the substrate.
  Retention is a declared policy with a meta-audit of deletions; a
  human does not run ad-hoc `DELETE` jobs.
  ```python
  # Fix: write to an append-only sink (illustrative — S3 Object Lock,
  # WORM bucket, or a hash-chained ledger). The API has no update /
  # delete path; the storage refuses mutation.
  def write_event(event, actor, target=None, payload=None, request_id=None):
      record = {
          "timestamp": datetime.datetime.now(datetime.timezone.utc).isoformat(),
          "event": event,
          "actor": actor,            # the authenticated principal, not "admin"
          "target": target,
          "payload": redact(payload), # see Breach 2 fix
          "request_id": request_id,
          "prev_hash": last_hash(),   # hash-chaining: tampering breaks the chain
      }
      record["hash"] = sha256(canonical_json(record))
      append_only_sink.write(record)  # WORM storage; no update/delete API exists
  # Fix: retention is a declared, reviewed policy — not an ad-hoc DELETE.
  # When an audit segment ages out, the deletion is itself meta-audited
  # in a higher-tier log with the rule that authorized it.
  # (See domains/compliance/data-retention.md and audit-logs.md.)
  ```
 - See `domains/compliance/audit-logs.md` (Audit Logs are Append-Only)
  and `domains/compliance/first-principles.md` P1.
 ### Breach 2 — Secret Leaked in Audit Log (Compliance P9 Secrets and Sensitive Data are Redacted in Audit)
 - `payload={"value": value}` writes the raw DB password into the
  audit record. The append-only log is now a secret store: anyone
  with `SELECT` on `audit_log` can read production credentials. The
  log is harder to secure than the secrets manager it read from.
 - Once the secret is in an append-only log, the remediation is
  expensive — rotate the secret *and* rewrite the log's access scope
  (you cannot edit the record; it is append-only). Redaction must
  happen *at the logging boundary, before the record is written*, not
  by opportunistic scrubbing after the fact.
 - The redaction policy here is "nothing" — there is no rule for
  which fields are redacted, by what mechanism, in which event type.
  A redaction rule that lives in no one's head and no code is a P9
  violation waiting to happen (and it happened).
 - **Fix:** redaction is structural, applied at the logging boundary
  before the record reaches the append-only sink. The policy is
  itself auditable (which fields, by what rule, in which event).
  ```python
  # Fix: redaction at the boundary. Log the FACT of the action
  # (a secret was read), never the CONTENT of the secret.
  REDACTED_FIELDS = {"value", "token", "password", "authorization", "secret"}
  def redact(payload):
      if not isinstance(payload, dict):
          return "[REDACTED:non-object]"
      out = {}
      for k, v in payload.items():
          if k.lower() in REDACTED_FIELDS or "secret" in k.lower():
              out[k] = "[REDACTED:secret]"
          else:
              out[k] = v
      out["_redaction"] = "secret-value-policy/v1"  # the rule is auditable
      return out
  # The fixed audit record:
  # {
  #   "event": "config.read",
  #   "actor": "api-server",            # the authenticated principal
  #   "target": {"kind": "secret", "id": "db-password"},
  #   "payload": {"value": "[REDACTED:secret]"},
  #   "_redaction": "secret-value-policy/v1",
  #   "request_id": "req_91c2"
  # }
  # The fact of the read is logged; the secret never enters the log.
  ```
 - See `domains/compliance/audit-logs.md` (Redaction at the Boundary)
  and `domains/compliance/first-principles.md` P9. Cross
  `domains/security/secrets.md` — the audit-side redaction is the
  complement of secret management.
 ## The Cascade (Two Breaches Compound)
 The two violations compound destructively. The secret enters the
 mutable log (P9 breach), where it sits readable by any `SELECT`-holder
 for a week. Then the `cleanup` job `DELETE`s the record (P1 breach) —
 destroying the evidence that the secret was ever logged, while the
 secret itself has already been exposed to every reader of the table.
 The `UPDATE` that rewrites `svc-deploy` → `admin` (a P7 attribution
 breach stacked on the P1 breach) means that even if a copy of the
 record survived, the actor who triggered the secret read is no longer
 identifiable. The team cannot answer "who read the DB password and
 when" — the log that would answer it was mutated, and the secret it
 leaked is now in the wild. This is the worst-case interaction of P1
 and P9: a secret leak with no attributable actor and no surviving
 record.
 ## Cross-Domain Links
 - `domains/compliance/audit-logs.md` — the append-only guarantee and
  the redaction-at-boundary rule this code violates.
 - `domains/compliance/first-principles.md` — P1 (Append-Only) and P9
  (Redacted) are the two breached principles; P7 (Attributable) is
  breached by the `UPDATE` rewrite.
 - `domains/compliance/evidence.md` — an audit log that can be
  `DELETE`d is not admissible evidence; the append-only guarantee is
  what makes it admissible.
 - `domains/compliance/data-retention.md` — retention is a declared
  policy with meta-audited deletions, not an ad-hoc `DELETE` job.
 - `domains/security/secrets.md` — redaction at the logging boundary
  is the audit-side complement of secret management.
 - `domains/observability/logging.md` — audit logs are structured
  logging with an append-only guarantee; the logging primitives
  compose here.
@@ -0,0 +1,175 @@
 # Bad Example: i18n String Concatenation
 > A checkout component that violates Atelier's i18n principles. Each
 > violation is cited, then fixed.
 ## The Code
 ```typescript
 // Checkout.tsx — the cardinal i18n violation
 function CartSummary({ itemCount, name, total, currency, date }) {
  // P3 VIOLATION: inline string concatenation. The source-language
  // word order ("Welcome, {name}! You have {n} items") is baked into
  // code. Every locale with different word order is broken.
  const welcome = "Welcome, " + name + "!";
  // P4 VIOLATION: hand-rolled plural branching. `if (n === 1)` encodes
  // exactly English's one/other rule. Arabic (six categories), Russian
  // (three), Polish (three) are all wrong.
  const items =
    itemCount === 1 ? "1 item" : itemCount + " items";
  // P5 VIOLATION: hand-rolled currency + date formatting. "$" + total
  // is wrong for de-DE (symbol, grouping, placement). The date
  // `getMonth() + 1 + "/" + getDay()` is US-only (mm/dd/yyyy).
  const price = "$" + total.toFixed(2);
  const d = new Date(date);
  const dateStr = (d.getMonth() + 1) + "/" + d.getDate() + "/" + d.getFullYear();
  return (
    <div>
      <h1>{welcome}</h1>
      <p>{items} · {price} · {dateStr}</p>
    </div>
  );
 }
 ```
 ```typescript
 // The "resource" file — a custom CSV the team hand-rolled.
 // locale,en_us
 // welcome_prefix,Welcome,
 // item_singular,item
 // item_plural,items
 //
 // This is a P3 violation on its own: a custom format no standard
 // tool (xgettext, i18next, FormatJS) can extract from or compose with.
 ```
 The team runs their first i18n test against real Arabic translations —
 after the string freeze, after the translator was paid. The Arabic
 build renders `"Welcome, محمد!"` with the name on the wrong side of
 the comma, `"1 items"` for a single item (Arabic has six plural
 categories, not two), and the price as `"$1,234.56"` (Arabic-Egypt
 formats as `"١٬٢٣٤٫٥٦ ج.م.‏"`). Every screen is a rewrite, not a patch.
 ## What Makes It Bad
 ### Inline String Concatenation (i18n P3 Resources are External, Not Inline)
 - `"Welcome, " + name + "!"` bakes English word order into code. In
  Japanese the name comes first (`ようこそ、محمدさん！`); in Arabic the
  structure differs again. The concatenation is invisible to the
  extraction pipeline (`xgettext`, `i18next-parser`) — the translator
  never sees it as a unit, and the string cannot be versioned or
  rolled back as a whole.
 - The custom `.csv` "resource" store is a second P3 violation: no
  standard tool reads it, it carries no plural grammar, and it cannot
  compose with the ICU formatting layer.
 - **Fix:** strings live in a standard locale resource file, addressed
  by key. Code calls `t("welcome", { name })`; the resource carries
  the parameterized message.
  ```json
  // en-US.json (ICU MessageFormat)
  {
    "checkout.welcome": "Welcome, {name}!",
    "checkout.cart.summary": "{count, plural, one {# item} other {# items}} · {price} · {date}"
  }
  ```
  ```json
  // ar-EG.json — six plural categories per CLDR; the code is identical
  {
    "checkout.welcome": "أهلاً بك، {name}!",
    "checkout.cart.summary": "{count, plural, zero {لا عناصر} one {عنصر واحد} two {عنصران} few {# عناصر} many {# عنصرًا} other {# عنصر}} · {price} · {date}"
  }
  ```
 - See `domains/i18n/locale-resources.md` (Resources are the Boundary)
  and `domains/i18n/first-principles.md` P3.
 ### Hand-Rolled Plural Branching (i18n P4 Plural and Gender are Parameterized)
 - `itemCount === 1 ? "1 item" : itemCount + " items"` encodes
  English's one/other rule and nothing else. Arabic has six
  categories (zero, one, two, few, many, other); Russian has three
  (one, few, many); Polish has three with different boundaries. A
  two-branch `if` is a C1 (Correctness) violation masquerading as a
  shortcut — it returns a wrong answer for every non-English locale.
 - **Fix:** the count goes to ICU MessageFormat; the formatter
  consults `Intl.PluralRules` for the active locale; the resource
  carries the variant for that category. The code passes the count,
  nothing more.
  ```typescript
  // The code passes the count; the resource + formatter pick the form.
  t("checkout.cart.summary", { count: itemCount, price, date });
  // Intl.PluralRules("ar-EG").select(1) === "one" -> "عنصر واحد"
  // Intl.PluralRules("ar-EG").select(2) === "two" -> "عنصران"
  // Intl.PluralRules("ar-EG").select(5) === "few" -> "٥ عناصر"
  ```
 - See `domains/i18n/locale-resources.md` (Plural and Gender in
  Resources) and `domains/i18n/formatting.md` (Plural Rules).
 ### Hand-Rolled Currency and Date Formatting (i18n P5 Formatting is Locale-Aware)
 - `"$" + total.toFixed(2)` hardcodes the US dollar symbol, US
  grouping (`,`), and US placement (symbol before the number). In
  `de-DE` the euro formats as `"1.234,56 €"` (symbol after, dot
  grouping). In `ar-EG` the pound formats as `"١٬٢٣٤٫٥٦ ج.م.‏"`
  (Arabic-Indic digits, different grouping).
 - `(d.getMonth() + 1) + "/" + d.getDate() + "/" + d.getFullYear()`
  produces `11/7/2024` — US `mm/dd/yyyy`. Most of the world reads
  `dd/mm/yyyy`; ISO is `yyyy-mm-dd`. A hand-rolled date formatter
  encodes one locale's convention and silently produces wrong output
  for every other.
 - **Fix:** `Intl.NumberFormat` and `Intl.DateTimeFormat` with a BCP
  47 locale tag. CLDR is the source of truth; `Intl` is the runtime.
  ```typescript
  new Intl.NumberFormat("ar-EG", { style: "currency", currency: "EGP" })
    .format(1234.56); // "١٬٢٣٤٫٥٦ ج.م.‏"
  new Intl.DateTimeFormat("ar-EG", { dateStyle: "medium" })
    .format(new Date(date)); // "٧ نوفمبر ٢٠٢٤"
  ```
 - See `domains/i18n/formatting.md` (the Intl surface, dates, numbers,
  currencies) and `domains/i18n/first-principles.md` P5.
 ### Source Language Treated as the Default (i18n P1 Source Language is a Locale)
 - The component has no resource layer at all for the source locale —
  English is "just the strings in the code." When the first second
  locale arrives, the fix is a rewrite (extract every string,
  restructure every concatenation), not a patch. The source language
  is `en-US`, a locale among many — it is not `null`.
 - **Fix:** extract source strings into `en-US.json` from day one,
  even before a second locale exists. The resource layer is the
  boundary from the first commit.
 - See `domains/i18n/first-principles.md` P1 and
  `domains/uiux/copywriting.md`.
 ## The Cascade
 The violations compound. Inline concatenation makes strings invisible
 to the extraction pipeline, so the translator never receives them as
 units — they reconstruct them by reading the code. Hand-rolled
 plurals return wrong answers for every non-English locale, so the
 Arabic build ships `"1 items"` for a single item. Hand-rolled
 formatting produces US-shaped output everywhere, so the price and
 date are wrong for `de-DE`, `ar-EG`, `zh-Hans-CN`, and every other
 locale. And because the first i18n test ran against real translations
 (a P8 violation — pseudo-locales should have surfaced all of this
 while the fix was still cheap), the defects are found after the
 string freeze, after the translator was paid, and after the release
 date was promised. The fix is now a re-translation and a re-release,
 not a commit.
 ## Cross-Domain Links
 - `domains/i18n/locale-resources.md` — the resource layer this code
  lacks; the standard formats (`.po`, JSON, Fluent, ICU Resource
  Bundle) it should have used.
 - `domains/i18n/formatting.md` — the `Intl`/ICU/CLDR formatting this
  code should call instead of hand-rolling.
 - `domains/i18n/first-principles.md` — P3, P4, P5, and P8 (pseudo-
  locales test early).
 - `domains/uiux/copywriting.md` — copy lives in resources, not in
  code.
 - `domains/api/error-responses.md` — the same parameterized-message
  discipline applies to localized API errors.
@@ -0,0 +1,226 @@
 # Good Example: AI/ML Reproducible Training Run
 > A training run that follows Atelier's AI/ML principles. Each aspect
 > cites the principle it satisfies. Scope per D-023: this is
 > engineering discipline (reproducibility, versioning, lineage,
 > serving), **not** algorithm or model design — no architecture
 > choice, hyperparameter tuning, or model-family comparison appears
 > here.
 ## The Run
 A training run `2026-08-05T09:12:00Z#run-42` produces model
 `registry/payments-fraud@sha256:b5e1...aa0`. Every input that shaped
 the model is pinned, named, and recoverable; the eval was declared
 before training; the model is an addressed artifact in a registry;
 the rollback path names the prior model and the prior dataset.
 ### The Reproducibility Contract
 ```yaml
 # lineage/run-42.yaml — the lineage root, committed alongside the code
 run_id:        2026-08-05T09:12:00Z#run-42
 dataset:       s3://ml-data/train@sha256:7f3a...e21
 splits:        dvc.yaml@commit a1b2c4d
 code:          git@a1b2c4d
 config:        configs/train.yaml@commit a1b2c4d
 environment:   ghcr.io/org/train-img@sha256:9c2d...f88
 eval_spec:     configs/eval.yaml@commit a1b2c4d
 model_digest:  registry/payments-fraud@sha256:b5e1...aa0
 status:        passed  # eval gate passed -> eligible for promotion
 ```
 - Lose any line and the run is anecdote, not evidence. The record is
  the lineage root: a prediction cites the `model_digest`, which
  cites this `run_id`, which cites everything above.
 ### Data is Versioned (DVC, content-hashed)
 ```ini
 # dvc.yaml — the split config is versioned in git, the data in the
 # content-addressed object store. Both are pinned by commit + hash.
 stages:
  prepare:
    cmd: python src/prepare.py --input data/raw --out data/splits
    deps:
      - data/raw
      - src/prepare.py
    outs:
      - data/splits/train.parquet
      - data/splits/val.parquet
      - data/splits/test.parquet
    # The dataset hash (sha256:7f3a...e21) is recorded in the lineage
    # contract above. "s3://ml-data/latest" would be a P2 violation.
 ```
 ```bash
 # The dataset is pinned by content hash, not by a mutable path.
 $ dvc get s3://ml-data/train --rev sha256:7f3a...e21
 # The split is a deterministic function of (dataset version, split
 # config, random seed). Two runs on the same pinned inputs produce
 # the same splits.
 ```
 ### Code and Config are Versioned (git)
 ```yaml
 # configs/train.yaml@commit a1b2c4d — versioned with the code
 # (No algorithm/hyperparameter content is illustrated here — this is
 #  the engineering discipline of pinning the config, not the model
 #  design inside it. Per D-023, algorithm choice is out of scope.)
 seed: 42
 splits:
  train: data/splits/train.parquet
  val:   data/splits/val.parquet
  test:  data/splits/test.parquet   # held out, never touched by training
 ```
 ### Environment is Pinned (container digest)
 ```dockerfile
 # The training environment is an image addressed by digest, not :latest.
 # ghcr.io/org/train-img@sha256:9c2d...f88
 FROM python:3.11-slim
 # dependencies pinned in requirements.txt with hashes
 COPY requirements.txt .
 RUN pip install --no-cache-dir -r requirements.txt
 ```
 ```text
 # requirements.txt — pinned + hash-pinned (pip-compile / pip-audit)
 dvc==3.50.2 \
    --hash=sha256:1c8a...e7
 mlflow==2.16.0 \
    --hash=sha256:9b2f...a1
 # No unpinned ranges. A rerun pulls the exact same wheels.
 ```
 ### Evaluation is Defined Before Training (P4)
 ```yaml
 # configs/eval.yaml@commit a1b2c4d — committed BEFORE training runs.
 # The metrics, splits, and pass/fail thresholds are a-priori; they
 # are the contract the model must satisfy to leave the experiment.
 metrics:
  - name: precision_at_threshold
    threshold: ">= 0.92"
  - name: recall_at_threshold
    threshold: ">= 0.85"
  - name: false_positive_rate
    threshold: "<= 0.03"
 split: data/splits/test.parquet   # held out, never in training
 gate: all_metrics_pass            # AND of all thresholds; no cherry-pick
 # The eval schema equals the serving input contract (serving.md P8):
 # feature names, types, ranges match the production boundary exactly.
 ```
 - Metrics chosen after seeing scores would be a P4 violation: the eval
  would be rationalizing, not measuring. See
  `domains/ai-ml/model-evaluation.md`.
 ### The Model is a Versioned Artifact (MLflow registry)
 ```bash
 # After the eval gate passes, the model is registered as an immutable
 # artifact addressed by digest, then promoted by stage.
 $ mlflow models register \
    --name payments-fraud \
    --model-uri runs:/run-42/model \
    --description "run-42, dataset sha256:7f3a...e21, eval passed"
 # registry/payments-fraud@sha256:b5e1...aa0
 # Stages: None -> Staging -> Production. Promotion is a registry
 # operation, not a file copy. Never "latest".
 ```
 ### The Pipeline Composes (P9)
 ```text
 # The training flow is a pipeline with explicit stages and contracts,
 # not a notebook. Each stage has named inputs and named outputs.
 prepare(dataset@hash) -> split(dvc.yaml) -> train(config, env@digest)
  -> eval(eval.yaml, test@hash) -> [gate: pass] -> register(model@digest)
                                  |
                                  +-> [gate: fail] -> abort, no promote
 # A notebook in this path would be a P9 violation: implicit state,
 # human-dependent order, unreproducible.
 ```
 ## What Makes It Good
 ### Reproducibility is First Class (AI/ML P1, C1, C5)
 - data + code + config + environment are all pinned. A second
  engineer on a second laptop checks out commit `a1b2c4d`, pulls the
  dataset by hash, pulls the image by digest, and reproduces the run
  bit-for-bit. The run is reviewable because it is recreatable.
 - See `domains/ai-ml/first-principles.md` P1 and
  `domains/devops/first-principles.md` P1 Reproducibility.
 ### Data is Versioned, Not Just Code (AI/ML P2, C5, C7)
 - The dataset is `s3://ml-data/train@sha256:7f3a...e21`, not
  `s3://ml-data/latest`. A model trained on "the data" is a model
  trained on an unknown input — a C1 violation. DVC pins the data the
  way git pins the code.
 - See `domains/ai-ml/data-versioning.md` (dataset hashing, the DVC /
  Delta Lake / LakeFS comparison) and `domains/data/migrations.md`.
 ### Lineage is Traceable End-to-End (AI/ML P3, C7, C1)
 - prediction → model → run-42 → dataset → source. Every edge is
  named; no orphan model. A serving regression traces back to the
  exact dataset and code that built the model, which is how drift is
  diagnosed (data drift vs concept drift vs prediction drift).
 - See `domains/ai-ml/data-versioning.md` (lineage record) and
  `domains/observability/logging.md`.
 ### Evaluation Defined Before Training (AI/ML P4, C1, C2)
 - `eval.yaml` was committed before `train` ran. The gate is
  `all_metrics_pass`; a failing metric aborts promotion. Cherry-
  picking a metric post-hoc is a correctness violation — the eval
  would no longer measure the model.
 - See `domains/ai-ml/model-evaluation.md` (eval-as-a-gate) and
  `domains/testing/first-principles.md` (tests as specification).
 ### Models are Versioned Artifacts (AI/ML P5, C5, C6)
 - The model is `registry/payments-fraud@sha256:b5e1...aa0`, promoted
  Staging → Production. A serving endpoint that pulled `latest` would
  be serving an unknown model with no rollback. The registry is to
  models what a container registry is to images.
 - See `domains/ai-ml/serving.md` (the model is an addressed artifact)
  and `domains/devops/first-principles.md` P7 Immutability.
 ### Rollback Includes the Model (AI/ML P10, C5)
 - If production regresses, the rollback restores the prior model
  digest `registry/payments-fraud@sha256:a1c4...f09` AND the prior
  serving code. A rollback that redeploys old code but keeps the new
  model has not rolled back — the model was the thing that regressed.
 - See `domains/ai-ml/serving.md` (Rollback Includes the Model) and
  `domains/devops/first-principles.md` P4 Rollback First.
 ## What This Example Does NOT Do (And Why That's Good)
 - Does **not** reference the dataset by a mutable path —
  `s3://ml-data/latest` would be a P2 violation.
 - Does **not** choose metrics after seeing scores — that is a P4
  violation (rationalizing, not measuring).
 - Does **not** pull `latest` from the model registry — that is a P5
  violation (unknown model, no rollback).
 - Does **not** contain algorithm/architecture/hyperparameter content
  — per D-023, those are research choices, not engineering
  principles, and have no derivation in the core C-rules.
 - Does **not** run from a notebook — a notebook in the pipeline path
  is a P9 violation (implicit state, unreproducible).
 ## Cross-Domain Links
 - `domains/ai-ml/data-versioning.md` — the DVC pinning, the lineage
  record, the tool comparison (DVC / Delta Lake / LakeFS).
 - `domains/ai-ml/serving.md` — the model is promoted as an addressed
  artifact; the serving boundary validates inputs against the same
  schema as the eval.
 - `domains/ai-ml/model-evaluation.md` — the eval-as-a-gate that this
  run must pass before promotion.
 - `domains/devops/first-principles.md` P1 Reproducibility — the
  non-negotiable this run inherits.
 - `domains/data/migrations.md` — data versioning parallels schema
  migration discipline.
 - `domains/observability/logging.md` — the lineage record is a
  structured, append-only log of provenance.
@@ -0,0 +1,197 @@
 # Good Example: GitOps Pull Request
 > A pull request that changes ArgoCD Application manifests following
 > Atelier's GitOps + Operators principles. Each aspect cites the
 > principle it satisfies.
 ## The PR
 A PR titled `promote payments-api 1.2.3 to prod` opened against the
 GitOps repo `platform/gitops`. It changes the `targetRevision` of the
 payments-api Application from `1.2.2` to `1.2.3`, adds a sync-wave
 annotation to a new migration Job, and tightens the AppProject's
 destination allow-list. CI runs plan/diff; nothing pushes to the
 cluster.
 ### The Commit
 ```yaml
 # manifests/prod/payments-api.yaml — the only file changed
 apiVersion: argoproj.io/v1alpha1
 kind: Application
 metadata:
  name: payments-api
  namespace: argocd
  finalizers:
    - resources-finalizer.argocd.argoproj.io
 spec:
  source:
    repoURL: https://git.example.com/platform/payments
    targetRevision: 1.2.3          # was 1.2.2 — pinned, not latest
    path: manifests/prod
  destination:
    server: https://kubernetes.default.svc
    namespace: payments
  syncPolicy:
    automated:
      prune: true
      selfHeal: true
    syncOptions:
      - CreateNamespace=false
      - PrunePropagationPolicy=foreground
 ```
 ```yaml
 # manifests/prod/payments-db-migration.yaml — new file, wave-ordered
 apiVersion: batch/v1
 kind: Job
 metadata:
  name: payments-db-migrate-1.2.3
  namespace: payments
  annotations:
    argocd.argoproj.io/sync-wave: "-1"   # PreSync: run before the app
 spec:
  backoffLimit: 0
  ttlSecondsAfterFinished: 86400
  template:
    spec:
      restartPolicy: OnFailure
      serviceAccountName: payments-migrator
      containers:
        - name: migrate
          image: registry.example.com/payments/migrate:1.2.3
          args: ["up", "--to", "1.2.3"]
 ```
 ```yaml
 # projects/payments.yaml — AppProject tightened in the same PR
 apiVersion: argoproj.io/v1alpha1
 kind: AppProject
 metadata:
  name: payments
  namespace: argocd
 spec:
  sourceRepos:
    - https://git.example.com/platform/payments
  destinations:
    - server: https://kubernetes.default.svc
      namespace: payments            # prod-app can only deploy to payments ns
  clusterResourceWhitelist:
    - group: ""
      kind: Namespace                # allowed to create its own namespace
  roles:
    - name: payments-team
      policies:
        - p, proj:payments:payments-team, applications, sync, payments/*, allow
 ```
 ### The CI Pipeline (runs on the PR, before merge)
 ```text
 # .github/workflows/gitops-plan.yml (illustrative steps)
 - name: validate manifests
  run: argocd app manifests manifests/prod/ | kubeconform -strict
 - name: diff against live cluster (read-only, no apply)
  run: argocd app diff payments-api --server $ARGOCD_SERVER --auth-token $READ_ONLY_TOKEN
  # CI holds a READ-ONLY ArgoCD token. It never holds kubectl rights.
  # A non-empty diff is the PR's proposed change, rendered for review.
 - name: opa gate (admission policy pre-check)
  run: opa eval -i manifests/prod/ -d policies/ "data.k8s.admission.deny"
  # Policy violations fail the PR before merge, not after deploy.
 ```
 ## What Makes It Good
 ### Git is the Source of Truth (GitOps P1, C1 Correctness)
 - The promotion is a commit. The cluster's desired state is a
  derivative of this repo; the repo is the authority. If the change is
  wrong, `git revert` is the rollback — the recovery path is the
  history.
 - See `domains/gitops-operators/first-principles.md` P1 and
  `domains/gitops-operators/argocd.md` (Application CRD).
 ### Pull, Don't Push (GitOps P3, C4 Locality)
 - CI holds a **read-only** ArgoCD token for `app diff`. It holds no
  `kubectl` rights against the production cluster. The cluster's
  ArgoCD controller pulls the merged commit; nothing pushes to the
  cluster. A compromised CI token can read, not deploy.
 - See `domains/gitops-operators/argocd.md` (RBAC and SSO) and
  `domains/gitops-operators/flux.md` for the same pull boundary from
  the Flux side.
 ### State is Immutable and Versioned (GitOps P5, C5 Reversibility)
 - `targetRevision: 1.2.3` — the Application pins a specific chart
  revision, not `latest`. The commit that changed it is a permanent
  record; `git revert` restores `1.2.2` and ArgoCD's `selfHeal`
  converges the cluster back. No force-push; history is the audit
  trail.
 - See `domains/gitops-operators/first-principles.md` P5 and
  `domains/infrastructure-as-code/state.md` (State is Truth).
 ### Sync Waves Order Correctness (GitOps P4, C1)
 - The migration Job carries `argocd.argoproj.io/sync-wave: "-1"` so
  it runs in `PreSync` before the payments-api Deployment that
  depends on the new schema. Wave ordering is a correctness
  mechanism, not performance — the app starting before its migration
  is a correctness bug.
 - See `domains/gitops-operators/argocd.md` (Sync Waves and Hooks).
 ### Reconcile, Don't Mutate by Hand (GitOps P8)
 - `selfHeal: true` + `prune: true` means a hand-edited drift on a
  managed resource is overwritten on the next loop. The fix for drift
  is a new commit, not `kubectl edit`. The PR author does not SSH into
  the cluster to "fix" anything.
 - See `domains/gitops-operators/argocd.md` (Diff and Drift) and
  `domains/gitops-operators/first-principles.md` P8.
 ### Least Privilege Reconciliation (GitOps P10, C8 Economy)
 - The AppProject `payments` restricts the Application to the
  `payments` namespace and the `payments` repo. The controller's
  ServiceAccount (not shown) is bound to a namespace-scoped Role, not
  `cluster-admin`. The PR *tightens* the allow-list — least privilege
  is a direction, not a one-time setting.
 - See `domains/gitops-operators/argocd.md` (RBAC and SSO) and
  `domains/kubernetes/rbac.md`.
 ### Policy is a Gate (Compliance P5, cross-link)
 - The `opa eval` step runs the admission policy against the proposed
  manifests before merge. A violation fails the PR; the non-compliant
  state is never realized. Detection is not enforcement; this is
  enforcement.
 - See `domains/compliance/policy-as-code.md` and
  `domains/devops/ci-cd.md`.
 ### Failure is Observable (GitOps P9)
 - A sync failure or health degradation on `payments-api` emits
  ArgoCD status (`Degraded` / `OutOfSync`) and a notification. Silent
  drift is the bug; this PR does not disable notifications.
 - See `domains/gitops-operators/argocd.md` (Health and Status) and
  `domains/observability/metrics.md`.
 ## What This PR Does NOT Do (And Why That's Good)
 - Does **not** run `kubectl apply` from CI — that is the push pattern,
  a P3 violation (see `examples/bad/` for the anti-pattern).
 - Does **not** use `argocd app set` as the steady state — the change
  is in git, not in an imperative command's history.
 - Does **not** store raw Secrets in the GitOps repo — secrets arrive
  via Sealed Secrets / SOPS / External Secrets, encrypted in git.
 - Does **not** float `targetRevision: latest` — the Application pins
  a version; "latest" is an unknown model of the system.
 ## Cross-Domain Links
 - `domains/gitops-operators/argocd.md` — the Application CRD, sync
  waves, RBAC/AppProjects, and the pull model.
 - `domains/gitops-operators/flux.md` — the same PR pattern from the
  Flux side (Kustomization CRD, per-cluster autonomy).
 - `domains/kubernetes/workloads.md` — the Deployment/Job the
  Application reconciles.
 - `domains/kubernetes/rbac.md` — the ServiceAccount + Role the
  controller and the migration Job run as.
 - `domains/compliance/policy-as-code.md` — the OPA gate is a
  compliance-as-a-gate enforcement point.
 - `domains/devops/P4 Rollback First` — `git revert` is the rollback;
  `selfHeal` is the convergence.
@@ -6,14 +6,14 @@
 | Core Principle | Domains that derive from it | Count |
 |----------------|---------------------------|-------|
-| C1 Correctness | All 13 (v0.1: 11; v0.2: infrastructure-as-code, kubernetes) | Universal |
+| C1 Correctness | All 17 (v0.1: 11; v0.2: infrastructure-as-code, kubernetes; v0.3: gitops-operators, ai-ml, i18n, compliance) | Universal |
-| C2 Clarity | v0.1: uiux, api, data, testing, observability, errors, documentation, devops; v0.2: infrastructure-as-code, kubernetes | 10 |
+| C2 Clarity | v0.1: uiux, api, data, testing, observability, errors, documentation, devops; v0.2: infrastructure-as-code, kubernetes; v0.3: gitops-operators, ai-ml, i18n, compliance | 14 |
-| C3 Simplicity | v0.1: security, data, testing, performance, documentation, concurrency, devops | 7 |
+| C3 Simplicity | v0.1: security, data, testing, performance, documentation, concurrency, devops; v0.2: infrastructure-as-code; v0.3: gitops-operators, i18n, compliance | 11 |
-| C4 Locality | v0.1: testing, concurrency; v0.2: infrastructure-as-code, kubernetes | 4 |
+| C4 Locality | v0.1: testing, concurrency; v0.2: infrastructure-as-code, kubernetes; v0.3: gitops-operators, i18n | 6 |
-| C5 Reversibility | v0.1: api, data, uiux, concurrency, devops; v0.2: infrastructure-as-code, kubernetes | 7 |
+| C5 Reversibility | v0.1: api, data, uiux, concurrency, devops; v0.2: infrastructure-as-code, kubernetes; v0.3: gitops-operators, ai-ml, i18n, compliance | 11 |
-| C6 Composability | v0.1: api, security, observability, errors, documentation, concurrency; v0.2: infrastructure-as-code, kubernetes | 8 |
+| C6 Composability | v0.1: api, security, observability, errors, documentation, concurrency; v0.2: infrastructure-as-code, kubernetes; v0.3: gitops-operators, ai-ml, i18n, compliance | 12 |
-| C7 Observability | v0.1: api, data, testing, performance, observability, errors, devops; v0.2: infrastructure-as-code, kubernetes | 9 |
+| C7 Observability | v0.1: api, data, testing, performance, observability, errors, devops; v0.2: infrastructure-as-code, kubernetes; v0.3: gitops-operators, ai-ml, i18n, compliance | 13 |
-| C8 Economy | v0.1: security, testing, performance, observability, concurrency; v0.2: kubernetes | 6 |
+| C8 Economy | v0.1: security, testing, performance, observability, concurrency; v0.2: kubernetes; v0.3: gitops-operators, i18n, compliance | 9 |
 ## Interpretation
@@ -34,11 +34,15 @@
 | Performance | C1, C3, C6, C7, C8 | Economy + measurement |
 | Observability | C1, C2, C6, C7, C8 | Self-referential (P7 derives from C7) |
 | Errors | C1, C2, C5, C6, C7 | Errors as data + reversibility |
-| Documentation | C1, C2, C3, C5, C6 | Docs-as-code + living docs |
+| Documentation | C1, C2, C3, C5, C6, C8 | Docs-as-code + living docs; economy of process (P1) |
 | Concurrency | C1, C3, C4, C5, C6, C7, C8 | Broadest derivation; locality (C4) unique |
-| DevOps | C1, C2, C3, C5, C7 | Reproducibility + rollback |
+| DevOps | C1, C2, C3, C5, C7, C8 | Reproducibility + rollback; economy of effort (P2) |
 | Infrastructure as Code | C1, C2, C3, C4, C5, C6, C7 | Declarative + state + composition; broadest derivation alongside Concurrency |
 | Kubernetes | C1, C2, C4, C5, C6, C7, C8 | Declarative + reversibility + economy; broad derivation (7 C-rules) |
 | GitOps + Operators | C1, C2, C3, C4, C5, C6, C7, C8 | Source-of-truth + reconciliation + pull-locality + least privilege; broadest derivation (8 C-rules, tied with i18n) |
 | AI / ML | C1, C2, C5, C6, C7 | Reproducibility + lineage + serving observability |
 | i18n | C1, C2, C3, C4, C5, C6, C7, C8 | Locale + formatting + direction + reversibility; broadest derivation (8 C-rules) |
 | Compliance | C1, C2, C3, C5, C6, C7, C8 | Audit + policy-as-code + retention + posture |
 ## v0.2 Domain Coverage (per IDEATE-03 schema)
@@ -47,10 +51,21 @@
 | Infrastructure as Code | 10 | 4 (terraform, opentofu, state, modules) | ✓ | complete |
 | Kubernetes | 10 | 6 (workloads, networking, storage, rbac, helm, kustomize) | ✓ | complete |
 ## v0.3 Domain Coverage (per IDEATE-03 schema)
 | Domain | P-count | Derived-doc-count | Manifest-listed | Status |
 |--------|---------|-------------------|-----------------|--------|
 | GitOps + Operators | 10 | 4 (argocd, flux, operators, progressive-delivery) | ✓ | complete |
 | AI / ML | 10 | 4 (data-versioning, model-evaluation, serving, monitoring-drift) | ✓ | complete |
 | i18n | 10 | 4 (locale-resources, formatting, rtl-bidi, testing-i18n) | ✓ | complete |
 | Compliance | 10 | 4 (audit-logs, data-retention, policy-as-code, evidence) | ✓ | complete |
 ## Gaps and Notes
 - No domain derives from only one C-rule. The minimum is 4 (UI/UX: C1, C2, C3, C5, C7 — actually 5). Every domain is multi-rooted.
- **Concurrency**, **Infrastructure as Code**, and **Kubernetes** are tied for the broadest derivation (7 C-rules each) — these domains touch the most core concerns.
+- **Concurrency**, **Infrastructure as Code**, and **Kubernetes** are tied for the broadest derivation among v0.1/v0.2 platform domains (7 C-rules each) — these domains touch the most core concerns.
 - **GitOps + Operators** and **i18n** are tied for the single broadest-derivation domain overall (8 C-rules each: C1–C8). GitOps adds C4 (pull-credential locality) alongside its source-of-truth/reconciliation/least-privilege derivation; i18n touches correctness, clarity, simplicity, locality, reversibility, composability, observability, and economy (expansion accommodation). This is consistent with both domains' cross-cutting nature.
 - **UI/UX** and **API** are the most user-facing; they emphasize C2 (Clarity) heavily.
 - **Security** is the only domain with explicit non-tradeable declarations; this promotes 8 of its rules to C1-equivalent per `core/conflict-resolution.md` §6.
- **v0.2 expansion:** C4 (Locality) grew from 2 to 4 domains (added infrastructure-as-code state locality, kubernetes namespace blast-radius). C6 (Composability) grew from 6 to 8. The two new domains are broad-derivation domains (7 C-rules each), consistent with Concurrency's breadth.
+- **v0.2 expansion:** C4 (Locality) grew from 2 to 4 domains (added infrastructure-as-code state locality, kubernetes namespace blast-radius). C6 (Composability) grew from 6 to 8. The two new domains are broad-derivation domains (7 C-rules each), consistent with Concurrency's breadth.
 - **v0.3 expansion:** C3 (Simplicity) grew from 7 to 11 (added gitops-operators declarative simplicity, i18n flexible layout, compliance structural redaction). C4 (Locality) grew from 4 to 6 (added gitops-operators pull-credential locality, i18n resource/text-direction locality). C5 (Reversibility) grew from 7 to 11 (added all four v0.3 domains — gitops history, ai-ml reproducibility, i18n translation versioning, compliance append-only/retention). C6 (Composability) grew from 8 to 12. C7 (Observability) grew from 9 to 13 (added all four v0.3 domains — reconciliation, drift detection, format correctness, posture). C2 (Clarity) grew from 10 to 14. The v0.3 expansion broadens every non-universal C-rule's coverage, confirming the four new domains are cross-cutting and well-rooted.
@@ -205,8 +205,68 @@ C5=Reversibility · C6=Composability · C7=Observability · C8=Economy
 | P9 Config and Secrets Sep | C2   | Clarity of configuration              |
 | P10 Roll Forward, Roll Back | C5 | Reversibility of deploys            |
-## Coverage Summary (post-v0.2)
+## GitOps + Operators
- 13 domains (11 v0.1 + 2 v0.2: infrastructure-as-code, kubernetes)
+| GitOps Principle           | Core | Why                                   |
- 130 domain principles total (110 v0.1 + 20 v0.2)
+|---------------------------|------|---------------------------------------|
 | P1 Git is the Source of Truth | C1, C5 | Correctness of source; reversibility via history |
 | P2 Declarative Over Imperative | C2, C3 | Clarity of intent; simplicity of expression |
 | P3 Pull, Don't Push        | C1, C4 | Correctness via security; locality of credentials |
 | P4 Continuous Reconciliation | C7, C1 | Observability of drift; correctness of converge loop |
 | P5 State is Immutable and Versioned | C5 | Reversibility through history |
 | P6 Operators Encode Domain Knowledge | C6, C2 | Composability of expertise; clarity of operations |
 | P7 Progressive Delivery is Reversible | C5, C1 | Reversibility of promotion; correctness of abort |
 | P8 Reconcile, Don't Mutate by Hand | C1, C7 | Correctness of source of truth; observability of drift |
 | P9 Failure is Observable and Surfaced | C7 | Observability of sync/rollout health |
 | P10 Least Privilege Reconciliation | C1, C8 | Correctness via security; economy of trust |
 ## AI / ML
 | AI/ML Principle            | Core | Why                                   |
 |---------------------------|------|---------------------------------------|
 | P1 Reproducibility is the First Class | C1, C5 | Correctness of runs; reversibility of reproduction |
 | P2 Data is Versioned, Not Just Code | C5, C7 | Reversibility of data; observability of dataset lineage |
 | P3 Lineage is Traceable End-to-End | C7, C1 | Observability of predictions; correctness of provenance |
 | P4 Evaluation is Defined Before Training | C1, C2 | Correctness of metrics; clarity of thresholds |
 | P5 Models are Versioned Artifacts | C5, C6 | Reversibility of model rollbacks; composability of registry |
 | P6 Serving is Observable    | C7   | Observability of inference            |
 | P7 Drift is Expected and Detected | C7, C1 | Observability of drift; correctness of detection |
 | P8 Inference Inputs are Validated | C1 | Correctness at serving boundary      |
 | P9 Pipelines Compose, Notebooks Don't | C6, C2 | Composability of steps; clarity of contracts |
 | P10 Rollback Includes the Model | C5 | Reversibility at the model layer     |
 ## i18n
 | i18n Principle             | Core | Why                                   |
 |---------------------------|------|---------------------------------------|
 | P1 Source Language is a Locale, Not the Default | C2, C1 | Clarity; correctness of localization model |
 | P2 Locale Identifiers are Standardized | C2, C6 | Clarity; composability of BCP 47      |
 | P3 Resources are External, Not Inline | C4, C6 | Locality of strings; composability of resources |
 | P4 Plural and Gender are Parameterized | C1, C6 | Correctness across locales; composability of message format |
 | P5 Formatting is Locale-Aware | C1, C7 | Correctness of formats; observability of format correctness |
 | P6 Text Direction is a Layout Primitive | C1, C4 | Correctness of RTL/bidi; locality of direction |
 | P7 Layout Accommodates Expansion | C8, C3 | Economy of rework; simplicity of flexible layout |
 | P8 Pseudo-Locales Test Early | C7, C5 | Observability of bugs early; reversibility of finding late |
 | P9 Images and Icons are Cultural | C1, C2 | Correctness; clarity of cultural meaning |
 | P10 Translation is Reversible and Versioned | C5 | Reversibility of localization changes |
 ## Compliance
 | Compliance Principle       | Core | Why                                   |
 |---------------------------|------|---------------------------------------|
 | P1 Audit Logs are Append-Only | C1, C5 | Correctness of audit; reversibility of immutable record |
 | P2 Every Significant Action is Logged | C7, C1 | Observability of actions; correctness of audit set |
 | P3 Retention is Policy, Not Storage | C5, C8 | Reversibility of lifecycle; economy of storage |
 | P4 Policy is Code         | C6, C2 | Composability of policy; clarity of rules |
 | P5 Policy is Evaluated as a Gate | C1, C5 | Correctness of enforcement; reversibility of block |
 | P6 Evidence is Collected Continuously | C7, C3 | Observability of posture; simplicity of audit |
 | P7 Identity is Attributable | C1, C7 | Correctness of attribution; observability of subject |
 | P8 Subject Access is Honored | C1, C5 | Correctness of rights; reversibility of deletion/export |
 | P9 Secrets and Sensitive Data are Redacted in Audit | C1, C3 | Correctness via security; simplicity of structural redaction |
 | P10 Compliance Posture is Observable | C7, C1 | Observability of compliance; correctness of posture |
 ## Coverage Summary (post-v0.3)
 - 17 domains (11 v0.1 + 2 v0.2: infrastructure-as-code, kubernetes; 4 v0.3: gitops-operators, ai-ml, i18n, compliance)
 - 170 domain principles total (110 v0.1 + 20 v0.2 + 40 v0.3)
 - Every domain P-rule traces to ≥1 core C-rule (C1–C8). No orphans.
@@ -147,6 +147,56 @@ If the task touches a domain, run that domain's checklist:
 - [ ] Rollout history retained; rollback tested (P10)
 - [ ] Namespaces used to bound blast radius; not `default` in prod (P6)
 ### If GitOps + Operators (see `domains/gitops-operators/`)
 - [ ] Desired state lives in git, not in the cluster (P1)
 - [ ] Configuration is declarative, not imperative scripts (P2)
 - [ ] Reconciliation is pull-based; no external push credentials into the cluster (P3)
 - [ ] Reconciliation loop runs continuously; drift auto-corrected (P4)
 - [ ] Every change is a commit; history is the audit/rollback path (P5)
 - [ ] Operational knowledge encoded as CRDs/controllers, not runbooks (P6)
 - [ ] Progressive delivery (canary/blue-green) has a tested abort/rollback path (P7)
 - [ ] No manual `kubectl apply`/`kubectl edit` on GitOps-managed resources (P8)
 - [ ] Sync failures, health degradation, and rollout stalls emit status + notifications (P9)
 - [ ] Controller credentials scoped to reconciled namespaces/resources; no cluster-admin GitOps robot (P10)
 ### If AI / ML (see `domains/ai-ml/`)
 - [ ] Scope check: this is engineering discipline (data versioning, evaluation, serving, drift), NOT algorithm/model design (D-023) — reject algorithm-design content
 - [ ] Every training run is reproducible from pinned data + code + config + environment (P1)
 - [ ] Datasets, features, and splits are versioned artifacts with lineage; `git` alone is insufficient (P2)
 - [ ] Any deployed prediction traces back through model → training run → dataset → source (P3)
 - [ ] Metrics, splits, and thresholds declared a priori; no post-hoc metric cherry-picking (P4)
 - [ ] Models are pinned, immutable, registry-tracked artifacts; never "the latest" (P5)
 - [ ] Inference latency, throughput, input distributions, and prediction confidence are observed (P6)
 - [ ] Data drift, concept drift, and prediction drift are monitored; a drift signal is an incident (P7)
 - [ ] Inference inputs validated against the model's contract (schema, ranges, types); out-of-contract rejected (P8)
 - [ ] Training/serving flows are composable pipelines with explicit steps; notebooks not in production (P9)
 - [ ] Serving rollback restores the prior model artifact, not just the prior code (P10)
 ### If i18n (see `domains/i18n/`)
 - [ ] Source language treated as one locale among many, not the "neutral" default (P1)
 - [ ] Locale identifiers use BCP 47 tags; no ad-hoc locale codes (P2)
 - [ ] User-facing strings in locale resource files, not concatenated inline in code (P3)
 - [ ] Plural/gender/select use ICU MessageFormat (or equivalent); no `if (n == 1)` branching (P4)
 - [ ] Dates, times, numbers, currencies, units via ICU/CLDR/`Intl`; no hand-rolled formatters (P5)
 - [ ] RTL/bidi is a first-class layout concern; logical CSS properties (`start`/`end`) over physical (`left`/`right`) (P6)
 - [ ] Layouts accommodate translation expansion; no fixed pixel widths for text (P7)
 - [ ] Pseudo-locales (accented, lengthened, RTL-mirrored) used to test before real translations arrive (P8)
 - [ ] Icons, colors, and imagery reviewed for locale-sensitivity; no locale-bound symbols treated as universal (P9)
 - [ ] Resource files versioned; a bad translation is a rollback, not a hot-patch (P10)
 ### If Compliance (see `domains/compliance/`)
 - [ ] Scope check: framework-agnostic — no regulation-specific (GDPR/HIPAA/SOC2/PCI) content (D-024)
 - [ ] Audit records are immutable once written; deletion/mutation is itself an auditable incident (P1)
 - [ ] The set of auditable actions is defined a priori; "we forgot to log it" is a violation (P2)
 - [ ] Data lifetime is declared and enforced as policy; deletion at end-of-life is a feature (P3)
 - [ ] Compliance policy expressed in versioned, reviewable, testable code (OPA/Cedar/Kyverno/Sentinel), not spreadsheets/prose (P4)
 - [ ] Policy violations block before the action (admission/CI/CD-time), not after the audit (P5)
 - [ ] Evidence gathered as a byproduct of operation, not assembled manually at audit time (P6)
 - [ ] Every logged action traces to an authenticated principal; no shared/generic identities (P7)
 - [ ] Data-subject rights (access, export, deletion) are operations with defined contracts and audit trails (P8)
 - [ ] Audit logs do not leak secrets; redaction is structural, not opportunistic (P9)
 - [ ] System reports its own compliance state (drift from policy, open violations, retention status) (P10)
 ## Final Gate
 - [ ] Have I read the relevant domain's first-principles?
@@ -163,4 +163,65 @@ When you see a pattern listed here, it is a defect. Cite the principle it violat
 |--------------|----------|-----|
 | Orphaned P-rule (a domain principle with no matrix row) | matrix completeness, C6 | Breaks the conflict-resolution arbiter; the rule has no core trace |
 | Deployable example artifact (standalone `.tf`/`.yaml` under `examples/`) | PROJECT.md "no runtime code", D-025 | Violates the docs-only contract; examples must be `.md` with fenced code |
-| Unlisted v0.2 doc (new doc not added to MANIFEST) | manifest rule | Not part of the framework by definition |
+| Unlisted v0.2 doc (new doc not added to MANIFEST) | manifest rule | Not part of the framework by definition |
 ## v0.3 Chaos Anti-Patterns (from IDEATE-20, IDEATE-24, IDEATE-25, IDEATE-27)
 These are named, cross-cutting violations specific to the v0.3 domains. Reject on sight.
 | Anti-Pattern | Breaches | Why |
 |--------------|----------|-----|
 | GitOps push-pattern (external CI pushes manifests to the cluster instead of an in-cluster agent pulling from git) | gitops P3 Pull, Don't Push; C1, C4 | Inverts the source-of-truth flow; requires push credentials into the cluster; breaks the reconciliation model (IDEATE-24, D-042) |
 | i18n LTR-only assumption (layout assumes left-to-right; no `dir` attribute, physical CSS properties only) | i18n P6 Text Direction is a Layout Primitive; C1, C4 | Disqualifying for RTL/Bidi users; locale-correctness violation (IDEATE-25, D-043) |
 | AI/ML orphan-model (a deployed prediction endpoint whose model has no lineage trace — no record of training run, dataset, or version) | ai-ml P3 Lineage is Traceable End-to-End; C7, C1 | Unreviewable, unrollbackable; the model is an unattributed artifact (IDEATE-27, D-045) |
 | Compliance mutable audit log (audit records can be edited or deleted by an operator) | compliance P1 Audit Logs are Append-Only; C1, C5 | Destroys the audit trail; the audit log's value is immutability — mutation is itself an incident |
 ### v0.3 Deployable Artifact Types (IDEATE-20, D-020)
 The following standalone file types are forbidden under `examples/` and elsewhere in the framework. Examples are `.md` files with fenced code only.
 | Forbidden standalone artifact | Belongs in | Why |
 |-------------------------------|-----------|-----|
 | `.po` / `.pot` resource files | fenced code in `examples/good/`/`examples/bad/*.md` | Runtime localization artifact; violates docs-only contract |
 | `.rego` / `.cedar` / `.sentinel` policy files | fenced code in `examples/*.md` | Runtime policy artifact; violates docs-only contract |
 | Model artifacts (`.pkl`, `.onnx`, `.pt`, `.h5`, `.safetensors`) | fenced code + prose in `examples/*.md` | Runtime model artifact; violates docs-only contract |
 | Signed manifests as standalone files (`.sig`, `.att`, `.intoto.jsonl`) | fenced code in `examples/*.md` | Runtime attestation artifact; violates docs-only contract |
 | Standalone `.yaml` / `.tf` / `.sh` | fenced code in `examples/*.md` | (Carried forward from v0.2) Runtime deployable artifact |
 ## v0.3 Domain-Specific Anti-Patterns
 ### GitOps + Operators
 | Anti-Pattern | Breaches | Why |
 |--------------|----------|-----|
 | Push-based deploy (external CI `kubectl apply` into the cluster) | P3 Pull, Don't Push | Inverts the model; requires push credentials; bypasses reconciliation |
 | Manual `kubectl apply`/`kubectl edit` on a GitOps-managed resource | P8 Reconcile, Don't Mutate by Hand | Unreconciled drift; the next loop overwrites it — silent and unattributed |
 | `cluster-admin` GitOps robot (controller bound to cluster-admin) | P10 Least Privilege Reconciliation | Overbroad grant; blast radius = entire cluster |
 | No sync-failure notification (silent drift on health degradation) | P9 Failure is Observable and Surfaced | Silent drift is the bug the loop was supposed to surface |
 ### AI / ML
 | Anti-Pattern | Breaches | Why |
 |--------------|----------|-----|
 | Unreproducible training run (unpinned data, code, config, or environment) | P1 Reproducibility is the First Class | Unreviewable; cannot debug, cannot rollback |
 | "Use the latest model" (serving points at `model:latest` instead of a pinned version) | P5 Models are Versioned Artifacts | Unversioned drift; rollback undefined |
 | Notebook in production (training/serving flow is a Jupyter notebook) | P9 Pipelines Compose, Notebooks Don't | No contracts, no composition, no reproducibility |
 | Orphan model (deployed prediction with no lineage trace) | P3 Lineage is Traceable End-to-End | Unattributed artifact; cannot trace to data/code (IDEATE-27) |
 ### i18n
 | Anti-Pattern | Breaches | Why |
 |--------------|----------|-----|
 | Inline string concatenation (`"Hello, " + name + "!"` in code) | P3 Resources are External, Not Inline | Not extractable; breaks translations; word-order differs per locale |
 | `if (n == 1)` plural branching (hand-rolled plural logic) | P4 Plural and Gender are Parameterized | Wrong for Arabic, Russian, Polish; ICU MessageFormat handles plurals |
 | LTR-only layout (no `dir` attribute, physical CSS `left`/`right`) | P6 Text Direction is a Layout Primitive | Disqualifying for RTL/Bidi (IDEATE-25) |
 | Hand-rolled date/number formatter (`new Date().toString()`, manual string formatting) | P5 Formatting is Locale-Aware | Locale-incorrect; ignores ICU/CLDR |
 ### Compliance
 | Anti-Pattern | Breaches | Why |
 |--------------|----------|-----|
 | Mutable audit log (operator can `UPDATE`/`DELETE` audit records) | P1 Audit Logs are Append-Only | Destroys the audit trail; mutation is itself an incident |
 | Shared/generic identity in audit (`admin` or `system` as the actor for all actions) | P7 Identity is Attributable | No attribution; no accountability; cannot investigate |
 | Secret leaked in audit log (request body or token captured in an audit event) | P9 Secrets and Sensitive Data are Redacted in Audit | Audit log becomes a secret exfiltration channel |
 | Manual evidence assembly at audit time (scramble to collect logs/scans/attestations on demand) | P6 Evidence is Collected Continuously | Audit-unready; evidence gathered under pressure is incomplete and unreliable |
@@ -79,6 +79,56 @@ Run the relevant domain section from `agent-checklist.md` (UI/UX, API, Security,
 - [ ] Are ConfigMaps and Secrets separate?
 - [ ] Is the rollback path tested, not assumed?
 ### If GitOps + Operators
 - [ ] Is desired state sourced from git, not from the cluster?
 - [ ] Is configuration declarative, not imperative scripts?
 - [ ] Is reconciliation pull-based (no external push credentials into the cluster)?
 - [ ] Does the reconciliation loop run continuously and auto-correct drift?
 - [ ] Is every change a commit, with history as the audit/rollback path?
 - [ ] Is operational knowledge encoded as CRDs/controllers, not runbooks humans must remember?
 - [ ] Does progressive delivery (canary/blue-green) have a tested abort/rollback path?
 - [ ] Are there manual `kubectl apply`/`kubectl edit` on GitOps-managed resources? (flag as incident)
 - [ ] Do sync failures, health degradation, and rollout stalls emit status + notifications?
 - [ ] Are controller credentials scoped to reconciled namespaces/resources (no cluster-admin GitOps robot)?
 ### If AI / ML
 - [ ] Scope check: is this engineering discipline (data versioning, evaluation, serving, drift), NOT algorithm/model design? (D-023 — reject algorithm-design content)
 - [ ] Is every training run reproducible from pinned data + code + config + environment?
 - [ ] Are datasets, features, and splits versioned artifacts with lineage (not just `git`)?
 - [ ] Can any deployed prediction trace back through model → training run → dataset → source?
 - [ ] Are metrics, splits, and thresholds declared a priori (no post-hoc metric cherry-picking)?
 - [ ] Are models pinned, immutable, registry-tracked artifacts (never "the latest")?
 - [ ] Is inference observable (latency, throughput, input distributions, prediction confidence)?
 - [ ] Are data drift, concept drift, and prediction drift monitored (drift signal = incident)?
 - [ ] Are inference inputs validated against the model's contract (schema, ranges, types)?
 - [ ] Are training/serving flows composable pipelines (not notebooks in production)?
 - [ ] Does serving rollback restore the prior model artifact, not just the prior code?
 ### If i18n
 - [ ] Is the source language treated as one locale among many, not the "neutral" default?
 - [ ] Do locale identifiers use BCP 47 tags (no ad-hoc locale codes)?
 - [ ] Are user-facing strings in locale resource files (not concatenated inline in code)?
 - [ ] Do plural/gender/select use ICU MessageFormat (no `if (n == 1)` branching)?
 - [ ] Are dates, times, numbers, currencies, units formatted via ICU/CLDR/`Intl` (no hand-rolled formatters)?
 - [ ] Is RTL/bidi a first-class layout concern (logical CSS properties over physical)?
 - [ ] Do layouts accommodate translation expansion (no fixed pixel widths for text)?
 - [ ] Are pseudo-locales used to test before real translations arrive?
 - [ ] Are icons, colors, and imagery reviewed for locale-sensitivity?
 - [ ] Are resource files versioned (bad translation = rollback, not hot-patch)?
 ### If Compliance
 - [ ] Scope check: is this framework-agnostic (no regulation-specific GDPR/HIPAA/SOC2/PCI content)? (D-024)
 - [ ] Are audit records immutable once written (deletion/mutation is itself an auditable incident)?
 - [ ] Is the set of auditable actions defined a priori ("we forgot to log it" is a violation)?
 - [ ] Is data lifetime declared and enforced as policy (deletion at end-of-life is a feature)?
 - [ ] Is compliance policy expressed in versioned, reviewable, testable code (not spreadsheets/prose)?
 - [ ] Do policy violations block before the action (admission/CI/CD-time, not after the audit)?
 - [ ] Is evidence gathered as a byproduct of operation (not assembled manually at audit time)?
 - [ ] Does every logged action trace to an authenticated principal (no shared/generic identities)?
 - [ ] Are data-subject rights (access, export, deletion) operations with defined contracts and audit trails?
 - [ ] Do audit logs avoid leaking secrets (redaction is structural, not opportunistic)?
 - [ ] Does the system report its own compliance state (drift from policy, open violations, retention status)?
 ## Review Etiquette
 - **Comment, don't command.** "This could be X" not "Change this to X."
Author	SHA1	Message	Date
Jon Chery	f61dffbb5a	docs(P02): complete messaging domain phase — v0.4 ---ci--- project: atelier phase: 2 milestone: v0.4 status: complete phase_role: execution phase_tag: v0.3.2 requirements: covered: [ATELIER-97, ATELIER-98, ATELIER-99, ATELIER-100, ATELIER-101] partial: [] ---/ci---	2026-08-05 16:00:31 +00:00
Jon Chery	f7f007dce8	docs(ship): P1 complete — v0.3.1 tagged, release 479 ---ci--- project: atelier phase: 1 milestone: v0.4 status: complete phase_role: execution phase_tag: v0.3.1 release_id: 479 ---/ci---	2026-08-05 15:53:16 +00:00
Jon Chery	20992883ff	docs(P01): complete edge domain phase — v0.4 ---ci--- project: atelier phase: 1 milestone: v0.4 status: complete phase_role: execution phase_tag: v0.3.1 requirements: covered: [ATELIER-92, ATELIER-93, ATELIER-94, ATELIER-95, ATELIER-96] partial: [] ---/ci---	2026-08-05 15:51:03 +00:00
Jon Chery	a622419e7d	docs(ship): P0 complete — v0.3.0 tagged, release 475 ---ci--- project: atelier phase: 0 milestone: v0.4 status: complete phase_role: pre_execution phase_tag: v0.3.0 release_id: 475 ---/ci---	2026-08-05 14:03:35 +00:00
Jon Chery	b1d293d218	docs(P00): complete pre-execution phase — v0.4 ---ci--- project: atelier phase: 0 milestone: v0.4 status: complete phase_role: pre_execution decisions: - id: D-054 decision: "v0.4 scope = edge + messaging + language-derived docs" - id: D-055 decision: "NFR milestone type — all phases docs/chore" - id: D-056 decision: "6 execution phases P1-P5 + final phase P6" - id: D-057 decision: "Edge domain groups CDN + offline-first + IoT + sync" - id: D-058 decision: "Messaging domain covers queues + pub/sub + streams + delivery-semantics" - id: D-059 decision: "Language-derived docs trace to existing domain P-rules, no new matrix rows" - id: D-060 decision: "4 derived docs per language, 16 total, per-language picks" - id: D-061..D-069 decision: "9 clarify resolutions (boundary, persona, examples, manifest, matrix)" - id: D-070..D-072 decision: "Research findings — edge P1-P10, messaging P1-P10, persona roster" - id: D-073 decision: "13 ideation recommendations accepted (12 refinements + 1 new req ATELIER-117)" - id: D-074..D-075 decision: "46-task vertical-slice plan, persona lifecycle" - id: G-013..G-018 decision: "Grill conditional proceed — G-014 false-premise fix applied" requirements: covered: [ATELIER-92, ATELIER-93, ATELIER-94, ATELIER-95, ATELIER-96, ATELIER-97, ATELIER-98, ATELIER-99, ATELIER-100, ATELIER-101, ATELIER-102, ATELIER-103, ATELIER-104, ATELIER-105, ATELIER-106, ATELIER-107, ATELIER-108, ATELIER-109, ATELIER-110, ATELIER-111, ATELIER-112, ATELIER-113, ATELIER-114, ATELIER-115, ATELIER-116, ATELIER-117] partial: [] ---/ci---	2026-08-05 14:01:13 +00:00
Jon Chery	9ebc9c8868	docs(milestone): complete v0.3 — GitOps+Operators/AI-ML/i18n/Compliance ---ci--- project: atelier phase: 6 milestone: v0.3 status: complete requirements: covered: [ATELIER-60..91] partial: [] ---/ci---	2026-08-05 03:45:38 +00:00