refactor(P21): rename acdl_platform/ -> core/ (REQ-53)

---ci---
project: acdl
phase: 21
milestone: v1.6
status: execute
---/ci---

Rename the acdl_platform/ package to core/ across the directory, all
imports in tests/scripts/pipelines/workflows, and doc references. The
package is imported as core.confidence_signal / core.contract_resolver /
core.outbox_writer. The deploy workflow's platform-repo checkout dir is
renamed acdl-platform/ -> platform/ (workspace path, not the python
package). Both .gitea + .github workflows stay byte-identical.

Note: the original target name 'platform/' shadows Python's stdlib
platform module (pytest's import uuid -> platform.system() fails when
the repo root is on sys.path, which every test does). 'core/' avoids
the clash while honoring the intent (drop the verbose acdl_platform).

Tests: 154 pass. run_ci.sh green.
This commit is contained in:
Jon Chery
2026-07-22 18:21:12 +00:00
parent c5745de37c
commit b758a7c242
22 changed files with 52 additions and 52 deletions
View File
View File
+103
View File
@@ -0,0 +1,103 @@
# ACDL Tiered Audit Ledger Design (REQ-20)
> **Status:** design authored in Phase 07 (milestone v1.1); the spike
> (Phases 08-10) implements the **v1.0 hash chain + DynamoDB outbox write**
> (D-041); the v1.2 build-out implements S3 Object Lock + JWS + async
> worker + DLQ + daily checkpoints.
The audit stream is the platform's tamper-evident record of every delivery
action. The vision's "Audit truth lives outside the repository" bet [1]
and "Not a mutable audit log" anti-goal [1] are the binding constraints.
Version-control history does not satisfy regulatory evidence; the ledger
is the source of truth.
## Three tiers
- **Cold tier (source of truth):** S3 with **Object Lock in compliance
mode**, **7-year retention** (ARCHITECTURE.md §9). No one — including
root — can delete or overwrite until retention expires. The regulatory
record.
- **Hot tier (query index):** the `acdl-evidence` audit repo (unchanged
from the v1.0 demo). Not part of the chain; a queryable mirror the
evidence UI (`evidence-ui/index.html`) reads. Lightweight attestation
linkage lives in the repo; the regulatory event body lives in S3.
- **Outbox (write path):** DynamoDB, **RPO = 0** (synchronous write before
contract submission ack). Single-region in v1 (`us-east-1`).
## Spike scope (D-041) — what Phases 08-10 implement
- **DynamoDB outbox:** table `acdl-outbox`, `PAY_PER_REQUEST` (D-044),
PK `contractId`, SK `eventType#eventTs`, TTL `expire_at` = now + 365d
(1-year storage per ARCHITECTURE.md §8).
- **`prev_event_hash` chain:** SHA-256 over canonical JSON
(`json.dumps(event, sort_keys=True, separators=(",", ":"))`), lifted
from the v1.0 demo's `evidence_writer.py`. Auto-genesis: first event
has `prev_hash="GENESIS"`.
- **Synchronous write** via boto3 `put_item` (strong-consistent by
default). No separate async worker / DLQ in the spike (RTO = workflow
re-run).
- **Mirror to `acdl-evidence`:** unchanged from v1.0 — the finalize step
commits `audit.json` to the evidence repo (the hot tier).
- **Spike evidence event shape:**
`{seq, ts, stage, event, prev_hash, hash, contractId, environment, stack, score, band}`.
## v1.2 build-out — what Phase 07 designs but the spike defers
- **S3 Object Lock:** bucket `acdl-evidence-lock-<account-id>`, Object
Lock enabled at creation, compliance mode, 7-yr retention
(`RetainUntilDate` = now + 7y). The outbox→S3 path is an async worker
that reads from the outbox and writes to Object Lock.
- **JWS detached signature (RFC 7515):** the event payload is
canonical-JSON-serialized, SHA-256 hashed, signed with a private key;
the signature is stored *detached* alongside the payload. Signing key =
**platform-level KMS key** (not per-contract — a per-contract key would
explode the key-management surface), rotated **quarterly**. The `jws`
field is added to the event shape in v1.2.
- **Async worker + DLQ:** a Lambda (or a Gitea Actions scheduled workflow)
reads the outbox, writes to S3 Object Lock, signs with KMS. DLQ = an
SQS dead-letter queue for failed writes. RTO = DLQ replay.
- **Daily checkpoints (§9):** a daily job reads the last event hash and
writes a "checkpoint" event to the ledger (+ optionally to a public
notarization service). The spike runs in minutes, not days — no
checkpoint in spike.
## JWS vs chain — orthogonality note
The `prev_event_hash` chain gives ordering/tamper-evidence *within* the
log (a deleted event breaks the chain visibly); JWS gives authenticity
*per event* (a forged event is detectable without re-reading the whole
chain). The chain is spike-scope; JWS is v1.2. Together they cover both
integrity properties the vision's "Not a mutable audit log" anti-goal
requires.
## Outbox item shape (full, spike + v1.2)
- PK `contractId` (UUID).
- SK `eventType#eventTs` (e.g. `POLICY_CHECKED#2026-07-21T12:00:00Z`).
- `payload` (the event body — hash-chained in spike, JWS-signed in v1.2).
- `prev_event_hash` (chain link; `GENESIS` for the first event).
- `hash` (this event's SHA-256 over canonical JSON).
- `approver_qa` (Gitea username of the QA approver; empty in dev-only
spike; populated on qa-promotion — D-042).
- `approver_prod` (SRE username; empty in spike).
- `environment`, `stack`, `score`, `band`.
- `expire_at` (TTL = now + 365d).
- **v1.2 only:** `jws` (detached signature), `checkpoint_ref`.
## RPO / RTO table
| Phase | RPO | RTO |
|-------|-----|-----|
| Spike (D-041) | 0 (sync outbox write) | workflow re-run |
| v1.2 | 0 (sync outbox) | async worker DLQ replay |
## Decision trail
- **D-041** — spike scope = hash chain + outbox write; Object Lock + JWS
+ worker + DLQ are v1.2.
- **D-044** — outbox mode `PAY_PER_REQUEST`; PK/SK; TTL `expire_at` =
now + 365d; no separate async worker in spike.
- **D-042** — approver identities (`approver_qa`, `approver_prod`) live
in the outbox; the separation-of-duties check
(`platform/separation_of_duties.py`) reads `approver_qa` and compares
to the prod-dispatch `gitea.actor`.
+175
View File
@@ -0,0 +1,175 @@
"""ACDL Confidence Signal (REQ-19).
The platform's certified answer to "is this safe to proceed?" (vision
tenet: "Safety is Computed, Not Assumed"). Every delivery action produces
a measurable, explainable confidence signal; reliance on operator
instinct is not a substitute.
Inputs (weights sum to 1.0, D-040):
1. policy_results (0.30) — list[PolicyCheckResult] (schemas/policy_check_result.schema.json)
2. validation (0.25) — {schema: bool, stack_resolved: bool, tf_validated: bool, tf_planned: bool}
3. freshness (0.10) — {age_days: float, max_age_days: float}
4. source (0.15) — {submitter: str, commit_sha: str, signed: bool}
5. history (0.10) — {prior_rollbacks: int, prior_policy_fails: int}
6. nfrs (0.10) — {declared: list[str], conformance: float|None}
Severity -> penalty (locked, ARCHITECTURE.md §8):
critical -> hard override (score = 0, block)
high -> -0.20
medium -> -0.05
low -> -0.01
info -> 0.00
Per-env thresholds (locked, ARCHITECTURE.md §8): dev 0.50, qa 0.75, prod 0.90, dr 0.95.
Output: {score, band, perInput, reasonCodes}.
Halt with explicit reason on missing input (§8).
Spike cold-start (A-6.2): inputs 3 (freshness), 5 (history), 6 (nfrs) are
'present + neutral 0.5' because the spike is the first submission with no
history and no declared NFRs. The gate is *presence*, not *conformance* —
the 'all six inputs present' dev gate (§5) is satisfied by non-null
per-input scores.
"""
from dataclasses import dataclass, asdict
from typing import List, Literal, Optional, Dict, Any
import json
import sys
WEIGHTS = {
"policy": 0.30,
"validation": 0.25,
"freshness": 0.10,
"source": 0.15,
"history": 0.10,
"nfrs": 0.10,
}
PENALTY = {
"critical": None,
"high": 0.20,
"medium": 0.05,
"low": 0.01,
"info": 0.0,
}
THRESHOLDS = {"dev": 0.50, "qa": 0.75, "prod": 0.90, "dr": 0.95}
@dataclass
class Signal:
score: float
band: Literal["pass", "warn", "block"]
perInput: Dict[str, float]
reasonCodes: List[str]
def _per_input_score(name: str, raw: Any) -> tuple:
"""Return (score in [0,1], reasons list). Unknown/missing -> 0.5 + INPUT_MISSING."""
reasons: List[str] = []
if raw is None:
return 0.5, [f"INPUT_MISSING:{name}"]
if name == "policy":
pcrs = raw if isinstance(raw, list) else []
if not pcrs:
return 0.5, []
scores = []
for pcr in pcrs:
r = pcr.get("result", "skipped")
if r == "pass" or r == "skipped":
scores.append(1.0)
else:
scores.append(0.0)
return sum(scores) / len(scores), []
if name == "validation":
keys = ("schema", "stack_resolved", "tf_validated", "tf_planned")
if not isinstance(raw, dict):
return 0.5, []
trues = sum(1 for k in keys if raw.get(k))
return trues / 4.0, []
if name == "freshness":
if not isinstance(raw, dict):
return 0.5, []
age = float(raw.get("age_days", 0))
mx = float(raw.get("max_age_days", 1)) or 1
s = 1.0 - (age / mx)
return max(0.0, min(1.0, s)), []
if name == "source":
if not isinstance(raw, dict):
return 0.5, []
if raw.get("submitter") and raw.get("commit_sha"):
return 1.0, []
return 0.5, []
if name == "history":
if not isinstance(raw, dict):
return 0.5, []
rollbacks = int(raw.get("prior_rollbacks", 0))
fails = int(raw.get("prior_policy_fails", 0))
s = 1.0 - (rollbacks * 0.2 + fails * 0.1)
return max(0.0, min(1.0, s)), []
if name == "nfrs":
if not isinstance(raw, dict):
return 0.5, []
conf = raw.get("conformance")
if conf is None:
return 0.5, []
return float(conf), []
return 0.5, []
def compute(contract_id: str, environment: str,
inputs: Dict[str, Any]) -> Signal:
"""Orchestrate the 6-input weighted sum + severity penalty + band."""
missing = sorted(set(WEIGHTS.keys()) - set(inputs.keys()))
if missing:
return Signal(0.0, "block", {},
[f"INPUT_MISSING:{m}" for m in missing])
per_input: Dict[str, float] = {}
reasons: List[str] = []
base = 0.0
for name, weight in WEIGHTS.items():
raw = inputs.get(name)
s, r = _per_input_score(name, raw)
per_input[name] = s
reasons.extend(r)
base += s * weight
penalty = 0.0
policy_input = inputs.get("policy")
pcrs = policy_input if isinstance(policy_input, list) else []
for pcr in pcrs:
if not isinstance(pcr, dict):
continue
if pcr.get("result") != "fail":
continue
sev = pcr.get("severity")
p = PENALTY.get(sev, 0.0)
if p is None:
return Signal(0.0, "block", per_input,
reasons + [f"CRITICAL_OVERRIDE:{pcr.get('ruleId','?')}"])
penalty += p
score = max(0.0, min(1.0, base - penalty))
threshold = THRESHOLDS[environment]
if score >= threshold:
band = "pass"
elif score < threshold - 0.10:
band = "block"
else:
band = "warn"
if environment == "dev" and band == "warn":
band = "block"
return Signal(score, band, per_input, reasons)
if __name__ == "__main__":
if len(sys.argv) < 3:
print("usage: confidence_signal.py <inputs.json> <environment>", file=sys.stderr)
sys.exit(2)
env = sys.argv[2]
with open(sys.argv[1], "r", encoding="utf-8") as fh:
inputs = json.load(fh)
sig = compute("cli", env, inputs)
print(json.dumps(asdict(sig), indent=2))
+262
View File
@@ -0,0 +1,262 @@
"""ACDL Contract Resolver — resolve a consumer contract to a Target Stack instance.
The contract resolver is the bridge between the consumer's declared intent
(a contract YAML) and the platform's executable representation (a Target
Stack JSON instance). It:
1. Loads and validates the contract against schemas/contract.schema.json.
2. Looks up the module name in modules/registry.json.
3. If the module is an L1 primitive: builds a stack instance directly from
the interface.json + contract inputs.
4. If the module is an L2 composition: loads the composition.json, expands
children to stack resources, resolves wires to ref: expressions, and
emits the full stack instance.
The output is a JSON instance valid against schemas/stack.schema.json,
ready for the Terraform adapter to compile.
CLI: contract_resolver.py <contract.yaml> <out.json>
"""
import json
import os
import sys
import yaml
import jsonschema
def _load_json(path):
with open(path, "r") as fh:
return json.load(fh)
def _load_yaml(path):
with open(path, "r") as fh:
return yaml.safe_load(fh)
def _resolve_wire_value(wire, contract_inputs, child_outputs):
"""Resolve a wire 'from' reference to a concrete value.
Wire 'from' can be:
- "contract.inputs.<name>" — a contract input value
- "<childId>.outputs.<name>" — a reference to another child's output
Returns either a concrete value (string/number/boolean) or a
"ref:<childId>.<outputName>" string for cross-child references.
"""
from_expr = wire["from"]
to_expr = wire["to"]
# If the 'from' is a contract input, use the concrete value
if from_expr.startswith("contract.inputs."):
input_name = from_expr[len("contract.inputs."):]
if input_name in contract_inputs:
return contract_inputs[input_name]
# Check for default
default = wire.get("default")
if default is not None:
return default
return None
# If the 'from' is a child output, emit a ref: expression
if "." in from_expr:
parts = from_expr.split(".", 2)
if len(parts) >= 3 and parts[1] == "outputs":
child_id = parts[0]
output_name = parts[2]
return f"ref:{child_id}.{output_name}"
return None
def resolve_l1(contract, registry, repo_root):
"""Resolve a contract referencing an L1 primitive to a stack instance."""
module_name = contract["module"]
module_ref = f"{module_name}@1.0.0"
inputs = contract.get("inputs", {})
environment = contract.get("environment", "dev")
# Load the interface
entry = registry[module_name]["1.0.0"]
iface_path = os.path.join(repo_root, entry["interface"])
iface = _load_json(iface_path)
# Build the stack instance
stack_instance = {
"version": "1.0.0",
"stack": {
"name": module_name,
"kind": "l1",
"depth": 1,
},
"resources": [
{
"id": iface.get("type", module_name).split(":")[-1]
if ":" in iface.get("type", "") else module_name,
"type": iface["type"],
"module": module_ref,
"inputs": dict(inputs),
"outputs": {
out_name: {"type": out_spec.get("type", "string")}
for out_name, out_spec in iface.get("outputs", {}).items()
},
}
],
}
# Add NFRs if present in the interface
nfrs = iface.get("nfrs", {})
if nfrs:
stack_instance["resources"][0]["nfrs"] = nfrs
return stack_instance
def resolve_l2(contract, registry, repo_root):
"""Resolve a contract referencing an L2 composition to a stack instance."""
module_name = contract["module"]
inputs = contract.get("inputs", {})
# Load the composition
entry = registry[module_name]["1.0.0"]
comp_path = os.path.join(repo_root, entry["interface"])
composition = _load_json(comp_path)
# Track child outputs for wire resolution
child_outputs = {}
resources = []
# Expand children to resources
for child in composition["children"]:
child_id = child["id"]
child_module = child["module"]
child_name = child_module.split("@")[0]
# Load the child's interface to get type and outputs
child_entry = registry[child_name]["1.0.0"]
child_iface_path = os.path.join(repo_root, child_entry["interface"])
child_iface = _load_json(child_iface_path)
# For multi-resource L1s (like vpc), the first resource type is the
# primary; the adapter handles expansion. Use the interface's type
# or the first resource in the interface's resources array.
if "resources" in child_iface and child_iface["resources"]:
# Multi-resource L1: create one resource per sub-resource
for sub_res in child_iface["resources"]:
resource = {
"id": f"{child_id}-{sub_res['type'].split(':')[-1].replace('_', '-')}"
if len(child_iface["resources"]) > 1 else child_id,
"type": sub_res["type"],
"module": child_module,
"inputs": {},
"outputs": {
out: {"type": "string"}
for out in sub_res.get("outputs", [])
},
}
resources.append(resource)
else:
# Single-resource L1
resource = {
"id": child_id,
"type": child_iface["type"],
"module": child_module,
"inputs": {},
"outputs": {
out_name: {"type": out_spec.get("type", "string")}
for out_name, out_spec in child_iface.get("outputs", {}).items()
},
}
resources.append(resource)
# Track outputs for this child
child_outputs[child_id] = child_iface.get("outputs", {})
# Resolve wires to populate inputs
for wire in composition.get("wires", []):
to_expr = wire["to"]
# Parse "to": "<childId>.inputs.<inputName>"
to_parts = to_expr.split(".")
if len(to_parts) != 3 or to_parts[1] != "inputs":
continue
target_child = to_parts[0]
input_name = to_parts[2]
value = _resolve_wire_value(wire, inputs, child_outputs)
if value is not None:
# Find the target resource and set the input
for res in resources:
if res["id"] == target_child or res["id"].startswith(f"{target_child}-"):
res["inputs"][input_name] = value
break
# Build the stack instance
stack_instance = {
"version": "1.0.0",
"stack": {
"name": module_name,
"kind": "l2",
"depth": composition.get("depth", 1),
},
"resources": resources,
}
return stack_instance
def resolve(contract_path, repo_root=None):
"""Resolve a consumer contract to a Target Stack instance.
Args:
contract_path: Path to the contract YAML file.
repo_root: Root of the ACDL repo (defaults to two levels up from this file).
Returns:
A dict representing the Target Stack instance.
"""
if repo_root is None:
repo_root = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
# Load contract
contract = _load_yaml(contract_path)
# Load schemas
contract_schema = _load_json(os.path.join(repo_root, "schemas", "contract.schema.json"))
# Validate contract against schema
jsonschema.validate(contract, contract_schema)
# Load registry
registry = _load_json(os.path.join(repo_root, "modules", "registry.json"))
module_name = contract["module"]
if module_name not in registry:
raise ValueError(f"module '{module_name}' not found in registry")
# Determine if L1 or L2
entry = registry[module_name]["1.0.0"]
interface_path = entry["interface"]
is_l2 = "l2" in interface_path or "composition" in interface_path
if is_l2:
stack_instance = resolve_l2(contract, registry, repo_root)
else:
stack_instance = resolve_l1(contract, registry, repo_root)
# Validate against stack schema
stack_schema = _load_json(os.path.join(repo_root, "schemas", "stack.schema.json"))
jsonschema.validate(stack_instance, stack_schema)
return stack_instance
if __name__ == "__main__":
if len(sys.argv) != 3:
print("usage: contract_resolver.py <contract.yaml> <out.json>", file=sys.stderr)
sys.exit(2)
result = resolve(sys.argv[1])
with open(sys.argv[2], "w") as fh:
json.dump(result, fh, indent=2)
print(f"resolver: resolved {sys.argv[1]} -> {sys.argv[2]}", file=sys.stderr)
+115
View File
@@ -0,0 +1,115 @@
# ACDL Human-in-the-Loop Matrix + Separation-of-Duties Design (REQ-21)
> **Status:** design authored in Phase 07 (milestone v1.1); v1.2 wires the
> gates. The spike (Phases 08-10) is **dev-only**; HITL is not exercised
> (the spike contract has `environment: dev`).
The vision's "Lower Environments are Autonomous; Higher Environments are
Attested" tenet [1] and the "deliberate human attestation — not as a
rubber stamp" requirement [1] are the binding constraints.
## Gate model (ARCHITECTURE.md §10.1)
**Pre-execution gates.** The contract is held in a "validated but not
applied" state until the human attests. qa, prod, dr are attestation
gates. No partial deployment to roll back on rejection (qa, prod); dr is
a separate deployment against a separate cluster/region. The
canary/deployment-rollback model is explicitly not in scope for v1.
## Gitea-specific gate mechanics (D-042)
Gitea has **no Environments API** and ignores `environment:` blocks
(v1.0 D-013; re-confirmed in RESEARCH TARGET 1). The pre-execution gate
is modeled as a `workflow_dispatch` with approval inputs:
- **qa gate:** `workflow_dispatch` with `approve_qa: true`; the dispatch
run's `gitea.actor` is the QA approver.
- **prod gate:** `workflow_dispatch` with `approve_prod: true`;
`gitea.actor` is the SRE approver.
- **dr gate:** `workflow_dispatch` with `approve_dr: true`; same.
The approver identity of record = `gitea.actor` of the dispatch run
(D-042). There is no other approval-identity signal in Gitea. The v1.2
real-OIDC path (blocked on go-gitea/gitea#36988) does not change this —
OIDC authorizes the *runner* to AWS, it does not change how the platform
records the *human* approver.
## Reviewer routing (ARCHITECTURE.md §10.2)
Gitea CODEOWNERS routes the right reviewer to the right gate:
- qa → QA team
- prod → SRE team
- dr → SRE team
CODEOWNERS **routes**; it does **not** enforce identity distinctness (that
is the platform-internal outbox check in
`platform/separation_of_duties.py`).
## Full 8-concern attestation matrix (§10.4, lifted verbatim)
| Env | Concern | Evidence artifact | Freshness | Source | Attester |
|---|---|---|---|---|---|
| qa | Functional correctness | Last successful run of contract-declared validation.e2eSuite with pass rate ≥ 99% | Last 24h | Test runner declared in contract | QA |
| qa | Performance baseline | Load test report (k6 / Gatling / Locust) showing p99 latency < declared NFR and throughput > declared minimum | Last 7d | Load test runner declared in contract | QA |
| qa | Security posture | Vulnerability scan (Trivy, Snyk, or contract-declared equivalent) with no criticals/highs, signed by Security on-call | Last 24h | Security scanner + Security team signature | QA |
| qa | Contract NFRs | Platform-generated report: schema valid, NFR assertions (latency, throughput, error rate) within declared bounds | At submission | Platform contract validator | QA |
| prod | Operational readiness | Runbook published, dashboard exists, on-call rotation assigned, alerts configured | At submission, validated against last 30d history | Platform + SRE | SRE |
| prod | Incident response | Sev-1 runbook tabletop or live drill completed | Last 90d | SRE drill record | SRE |
| prod | Capacity / cost | FinOps forecast for next 30d within budget envelope, cost anomaly baseline stored, budget alert configured | Forecast valid for next 30d | FinOps + SRE | SRE |
| prod | Resilience | DR drill, chaos engineering report, backup verified | DR: 180d; chaos: 90d; backup: 30d | SRE + Platform | SRE |
| dr | dr-region deploy with the most recent prod-bound dr drill as canary evidence | dr drill report | Last 180d | SRE | SRE |
## Timeout behavior (§10.5)
| Time | State | Action |
|---|---|---|
| Submission | PENDING_ATTESTATION | Notify responsible team |
| 1 business day | PENDING_ATTESTATION_WARNING | Notify team + platform on-call (elevated path); emit `PENDING_ATTESTATION_TIMEOUT_WARNING` event |
| 2 business days | PENDING_ATTESTATION_AUTO_FREEZE | Auto-freeze; require re-submission; emit `PENDING_ATTESTATION_AUTO_FREEZE` event; new submission linked via `supersedes` |
**Implementation:** a Gitea `on: schedule` workflow (runs hourly) that
scans the DynamoDB outbox for `PENDING_ATTESTATION` events with `ts`
older than 1/2 business days and emits the warn/freeze events. Not
implemented in the spike (dev-only).
## Rejection and rollback (§10.6)
Rejection returns the contract to a `HELD` state with the rejection
reason captured as a `PROMOTION_REJECTED` event. The consumer fixes the
cause and re-submits; the new submission is linked to the rejected one
via `supersedes` (a contract-schema field — `schemas/contract.schema.json`).
The audit chain is **extended, not torn up** (the "Not a mutable audit
log" anti-goal). No partial deployment to roll back at any v1 gate.
## Separation of duties (§10.3) — pointer to the .py
The identity-distinctness check is platform-internal, not GitHub-native,
not Kyverno (in v1). Sequence:
1. On promotion dev → qa, the platform reads the QA approver's identity
from the `workflow_dispatch` run's `gitea.actor` and writes it to the
DynamoDB outbox keyed by `contractId` (attribute `approver_qa`).
2. On promotion qa → prod, the platform reads the stored `approver_qa`
from the outbox and the new SRE approver's `gitea.actor` from the
prod-dispatch run.
3. If `approver_qa == approver_prod`, the platform blocks the prod
promotion, writes a `SEPARATION_OF_DUTIES_VIOLATION` event to the
evidence stream, and routes a halt artifact to the SRE on-call.
4. The check is implemented in `platform/separation_of_duties.py`
(T-7.8). The platform is the only writer to the outbox; the check is
in the same process that has authority to block the promotion.
## Spike scope note
The spike is dev-only (REQ-27 contract has `environment: dev`), so HITL
is not exercised. Phase 07 authors the design; Phase 10's
`verify_phase10.sh` does not assert HITL behavior. v1.2 wires the gates
against this design.
## Decision trail
- **D-042** — approver identity = `gitea.actor` of the `workflow_dispatch`
run; no Environments API in Gitea.
- **D-013** (v1.0) — the `workflow_dispatch` approval-input fallback,
re-used for the real platform's pre-execution gate model.
+71
View File
@@ -0,0 +1,71 @@
"""ACDL Outbox Writer — write an evidence event to the DynamoDB outbox.
ARCHITECTURE.md §9: DynamoDB outbox, RPO=0 (synchronous write before
ack). The event is hash-chained (SHA-256 over canonical JSON); the first
event has prev_event_hash="GENESIS". D-P10-3: the spike writes ONE
CONFIDENCE_COMPUTED event.
The outbox table (Phase 08): acdl-outbox, PAY_PER_REQUEST, PK contractId,
SK eventType#eventTs, TTL expire_at = now + 365d (D-044).
CLI: outbox_writer.py <event.json> (uses AWS creds from env)
"""
import datetime
import hashlib
import json
import os
import sys
import boto3
OUTBOX_TABLE = "acdl-outbox"
REGION = os.environ.get("AWS_DEFAULT_REGION", "us-east-1")
def _canonical_hash(event):
"""SHA-256 over canonical JSON (sort_keys, compact separators)."""
canonical = json.dumps(event, sort_keys=True, separators=(",", ":"))
return hashlib.sha256(canonical.encode("utf-8")).hexdigest()
def write_event(event, outbox_table=OUTBOX_TABLE, region=REGION):
"""Write an evidence event to the DynamoDB outbox. Returns the item dict."""
contract_id = event["contractId"]
event_type = event.get("eventType", "CONFIDENCE_COMPUTED")
event_ts = event.get("ts") or datetime.datetime.now(datetime.timezone.utc).strftime("%Y-%m-%dT%H:%M:%SZ")
sk = f"{event_type}#{event_ts}"
# Chain: first event = GENESIS (D-P10-3 spike writes one event).
prev_hash = event.get("prev_event_hash", "GENESIS")
event_hash = _canonical_hash(event)
item = {
"contractId": {"S": contract_id},
"eventType#eventTs": {"S": sk},
"payload": {"S": json.dumps(event, sort_keys=True)},
"prev_event_hash": {"S": prev_hash},
"hash": {"S": event_hash},
"environment": {"S": str(event.get("environment", ""))},
"stack": {"S": str(event.get("stack", ""))},
"score": {"N": str(event.get("score", 0))},
"band": {"S": str(event.get("band", ""))},
"expire_at": {"N": str(int((datetime.datetime.now(datetime.timezone.utc) +
datetime.timedelta(days=365)).timestamp()))},
}
session = boto3.Session(region_name=region)
dyn = session.client("dynamodb")
dyn.put_item(TableName=outbox_table, Item=item)
return item
if __name__ == "__main__":
if len(sys.argv) != 2:
print("usage: outbox_writer.py <event.json>", file=sys.stderr)
sys.exit(2)
with open(sys.argv[1], "r") as fh:
event = json.load(fh)
item = write_event(event)
print(json.dumps({k: list(v.values())[0] for k, v in item.items()}, indent=2))
+42
View File
@@ -0,0 +1,42 @@
"""Check that qaApprover != prodApprover for a contract (ARCHITECTURE.md
§10.3, D-042). Reads `approver_qa` from the DynamoDB outbox for the
contractId, compares to the prod-dispatch `gitea.actor`. Blocks on
equality, emits `SEPARATION_OF_DUTIES_VIOLATION`, routes a halt artifact
to SRE on-call.
Spike scope (A-8.1): the spike is dev-only (REQ-27 contract has
environment: dev); HITL is not exercised. This module is authored to its
full v1.2 shape but the spike calls it with current_prod_approver=None
and a None outbox_client — the check returns (True, 'no QA approver
recorded (dev-only spike)').
"""
from typing import Optional, Tuple
def check(outbox_client, contract_id: str,
current_prod_approver: Optional[str]) -> Tuple[bool, str]:
"""Return (ok, reason). ok=False means block the prod promotion."""
if outbox_client is None:
return (True, "no outbox client (dev-only spike)")
item = outbox_client.get(contract_id)
if item is None:
return (True, "no prior approver (first promotion)")
qa_approver = item.get("approver_qa")
if not qa_approver:
return (True, "no QA approver recorded (dev-only spike)")
if current_prod_approver is None:
return (True, "no prod approver supplied (dev-only spike)")
if qa_approver == current_prod_approver:
return (False,
f"SEPARATION_OF_DUTIES_VIOLATION: "
f"qaApprover==prodApprover=={qa_approver}")
return (True, "distinct")
def route_halt_artifact(contract_id: str, violation_reason: str,
oncall_client) -> None:
"""Route a halt artifact to SRE on-call. Spike: stub that logs. v1.2
wires a real pager."""
print(f"[halt-artifact] contract={contract_id} reason={violation_reason} "
f"oncall={oncall_client}", flush=True)