cloudinit-bot
c1aa274ab1
P5 complete. Bearers skeleton II (D-020/D-035 pattern, zero ext deps): - x/hub/types (NEW): HubService enum (3), LendingCouponCapBps=800 LOCAL const cross-documented to D-028 (A-304, no x/bond import), ClampLendingCoupon. - x/services/types (NEW): ServiceKind enum (4), window-id + operator-reach-id by-ID-string refs (A-307, G-003). - x/bond/types (EXT): GrowthBond, ClampGrowth with G-012 underflow guard (currentBps >= cap returns 0, no uint32 underflow), OrderSide/OrderStatus/ SecondaryOrder. D-028 regression (800/0 consts unchanged). Coverage: hub 93.3%, services 100%, bond 95.1%. Both firewalls green. G-003 intact. ---ci--- project: oy phase: 5 milestone: v0.3 status: complete tag_base: v0.2.x phase_role: execution requirements: covered: [REQ-024, REQ-025, REQ-026] partial: [] ---/ci---
OpenYield
OpenYield is a jurisdiction-light, public-good mesh for real production — a protocol organized around Holders, Stands, and the Six Principles, designed to hold real value without the words or the shapes that invite capture. The mesh runs on OY Chain (Layer 1), a canonical state layer for the Bread unit, the Storage Pools (Stash, Vault, Root-Pool), Standing, Watcher attestations, and the Pact / Council / Partner surface. It is anti-greed by construction: Mission Lock fixes the Six Principles and fee covenant so no council can amend them, and the 8% coupon cap on bonds is a mission-locked ceiling, not a parameter.
The Six Principles
- Real value — the mesh holds real production, not speculation.
- Sustainability — fees are floored and capped; the protocol cannot drain its users.
- Mission-lock — the Six Principles and fee covenant are immutable; no council can amend them.
- Openness — anyone may join; the mesh is a public good.
- Ownership — Holders own their Stash and their Reach; custody is theirs.
- Self-service — a Holder can act without a custodian; the mesh is jurisdiction-light.
Bread unit & scale
The unit of value is Bread, scaled in 11 tiers: Grain → Crumb → Bread → Loaf → Batch → Cake → Bakery → Granary → Mill → Harvest → Earth.
Status
v0.3 (Bearers & Documentation) — in progress. The codebase is a skeleton +
tests layer (Go types + keeper stubs + invariant tests, zero external Go deps)
matching the v0.1/v0.2 pre-MVP pattern. See .ciagent/oy/ROADMAP.md for the
phase plan and .ciagent/oy/PROJECT.md for governance.
Build & test
OpenYield is pure Go with zero external dependencies (go.mod has no
require lines; go 1.22). From the repo root:
go build ./...
go test ./...
Docs
The docs site is MkDocs Material
(a build-only Python dep; not a Go dep — go.mod is unchanged). To
preview locally:
mkdocs serve
# or build to a static site/ dir:
mkdocs build
The site lives under docs/ (see mkdocs.yml for the nav). Publishing CI is
deferred to v0.4 (D-046); v0.3 ships the source.
Lexicon firewall
OpenYield bans 10 financial terms as standalone words (REQ-012) across all Go
source (x/**/*.go) and all docs (README.md + docs/**/*.md). The banned
terms are the words you would expect a legacy financial institution to use;
this README and the docs describe them only by their safe replacements, so
the firewall itself never trips. The firewall is enforced in code by two
sibling Go tests:
lexicon_meta_test.go(v0.2) — scansx/**/*.go.lexicon_meta_docs/lexicon_meta_docs_test.go(v0.3) — scansREADME.md+docs/**/*.md.
Both use lexicon.FindBannedTerm (word-boundary, case-insensitive), so
"OpenYield" is safe (word-boundary does not match the banned term inside an
identifier) but the standalone banned term is not — docs say "real
production" / "real return", and a Holder's identity is Holder /
Reach, never the banned word for a custodial position. See
docs/shared/lexicon.md for the glossary of safe replacements.
Governance
.ciagent/oy/PROJECT.md— full vision, decisions (D-0xx), assumptions..ciagent/oy/PLANS.md— phase plans (v0.1, v0.2, v0.3)..ciagent/oy/REQUIREMENTS.md— REQ coverage matrix..ciagent/oy/ROADMAP.md— release roadmap.