v1.2.0 — add spec/plan/overview/registry types, ledger anchoring, examples, bugfixes#3
v1.2.0 — add spec/plan/overview/registry types, ledger anchoring, examples, bugfixes#3glandua wants to merge 1 commit into
Conversation
…g section, examples, bugfixes Version bump v1.1.0 → v1.2.0 (minor; no breaking changes). Additions - 4 new object types codified from concrete use-cases that had no clean home in v1.1.0: spec, plan, overview, registry. Each grounded in a RegenOS Phase 0 artifact serving as pilot evidence (CONTRIBUTING.md minimum-3 requirement satisfied with 4 pilots). - §Ledger Anchoring section: when to anchor, how to anchor (canonical/manifest/receipt triad via MsgAnchor), post-anchor rules, rename discipline classification procedure, §Legacy Names note grandfathering pre-v1.2.0 non-canonical on-chain anchors. - examples/ directory with one worked example per new type plus examples/README.md reading guide. - "Choosing Between Types" guidance extended with v1.2.0 pairs: docs vs spec, docs vs registry, strategy vs plan, readme vs overview. - Access Levels note clarifying that projects may define finer-grained sub-tiers without altering the canonical three-tier scheme. Fixes - docs/semantic-naming-properties.md: replace broken example across all three occurrences. Prior example had undefined type 'objective', double-version 'v2025-Q2.v0.1.4', and undefined status 'active'. Replaced with canonical core.strategy.regen-os-overview.v0.1.0. - CONTRIBUTING.md: restore missing ## prefix on "Who Can Contribute?" heading that was breaking Markdown rendering. Pilot evidence - core.spec.regen-os-learning-subroutines.v0.1.0 — atomic closed-loop pattern specification - core.plan.regen-os-phase-delivery.v0.1.0 — 5-phase execution plan - core.overview.regen-os-current-state.v0.1.0 — point-in-time explainer - core.registry.regen-os-artifact-registry.v0.1.0 — 45-artifact IRI/URL/RID index Legacy anchor case-study for §Ledger Anchoring: strategy.plan.claims-engine-dogfood-plan.v0.1.0 (mainnet, 2026-04-17) uses 'strategy' as first field, which is neither canonical relevance nor type. Grandfathered per §Legacy Names; preserved as historical record. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
There was a problem hiding this comment.
Code Review
This pull request updates the Knowledge Organization Infrastructure (KOI) naming convention to version 1.2.0. Key additions include four new object types—spec, plan, overview, and registry—along with a new section on ledger anchoring and rename discipline for anchored artifacts. The PR also introduces a directory of pilot examples and fixes several documentation inconsistencies. Review feedback correctly identified a naming mismatch between the manifesto examples and the pilot files, as well as a conceptual error in the legacy naming section regarding the classification of the "strategy" type.
| core.spec.regen-os-learning-subroutines.v0.1.0 | ||
| core.plan.regen-os-phase-delivery.v0.1.0 | ||
| core.registry.regen-os-artifact-registry.v0.1.0 | ||
| core.overview.regen-commons-current-state.v0.1.0 |
There was a problem hiding this comment.
The example core.overview.regen-commons-current-state.v0.1.0 is inconsistent with the pilot name used in the types table (line 83) and the actual pilot file in the examples/ directory (core.overview.regen-os-current-state.v0.1.0). Updating this for consistency ensures the manifesto accurately reflects the pilot set.
| core.overview.regen-commons-current-state.v0.1.0 | |
| core.overview.regen-os-current-state.v0.1.0 |
|
|
||
| ### Legacy Names | ||
|
|
||
| Some objects anchored prior to v1.2.0 ratification use a **non-canonical first field** (e.g., `strategy.plan.claims-engine-dogfood-plan.v0.1.0` where `strategy` is neither a canonical relevance nor type). These legacy names are preserved as historical record — the ledger anchor is the authoritative identity of the object, and the on-chain record cannot be retroactively corrected without invalidating the anchor. |
There was a problem hiding this comment.
The parenthetical note states that strategy is neither a canonical relevance nor type. However, strategy is explicitly defined as a canonical type in the v1.1.0 section (line 67). The issue with the legacy name is that a type was used in the relevance field, not that the term itself is unrecognized. The text should be corrected to avoid confusing readers about the validity of the strategy type.
| Some objects anchored prior to v1.2.0 ratification use a **non-canonical first field** (e.g., `strategy.plan.claims-engine-dogfood-plan.v0.1.0` where `strategy` is neither a canonical relevance nor type). These legacy names are preserved as historical record — the ledger anchor is the authoritative identity of the object, and the on-chain record cannot be retroactively corrected without invalidating the anchor. | |
| Some objects anchored prior to v1.2.0 ratification use a **non-canonical first field** (e.g., strategy.plan.claims-engine-dogfood-plan.v0.1.0 where strategy is a type, not a canonical relevance level). These legacy names are preserved as historical record — the ledger anchor is the authoritative identity of the object, and the on-chain record cannot be retroactively corrected without invalidating the anchor. |
Summary
Proposes KOI Naming Convention Manifesto v1.2.0 — a minor-version update (no breaking changes) addressing four concrete gaps observed while operationalizing RegenOS.
spec,plan,overview,registry— codified from real use-cases that had no clean home in v1.1.0examples/directory — one worked example per new type, serving as the CONTRIBUTING.md minimum-3-object pilot setdocs/semantic-naming-properties.mdbroken example,CONTRIBUTING.mdmissing heading prefixWhy Each Change
Each addition is grounded in a specific concrete artifact from today's RegenOS Phase 0 session (2026-04-21 builders-standup prep). No hallucinated practice — every claim traces to a file that now exists.
spectypeToday I needed to name an agentic loop specification defining three closed-loop patterns (GTM thesis loop, ai-ops-harness retro loop, engineering-product Spec Gate loop). No v1.1.0 type fit —
docsis too general (this is implementation-ready, not reference material);strategyis forward-looking framework (this is a spec to execute against).specfills the gap.plantypeThe 2026-04-17 mainnet-anchored artifact
strategy.plan.claims-engine-dogfood-plan.v0.1.0already usesplanas its type in practice, butplanwasn't in v1.1.0's type list (and incidentally that anchor uses a non-canonical first field — see §Legacy Names). Codifyingplanaligns the manifesto with existing anchored practice. Aplandiffers fromstrategy: strategy is the framework (why + what); plan is the execution sequence (when + who + gates).overviewtypeTop-level explainers of a system or domain at a point in time. Not a
readme(those are entry-point navigation for a directory); not astrategy(those are forward-looking); not amemo(those originate a position). Grounded in today'sregen-os-overview-v0.1.mdartifact.registrytypeMeta-documents whose primary content is the index itself — not the content of what's indexed. Grounded in today's 45-artifact IRI/URL/RID registry. Downstream agents use it as a lookup table. Updates follow an add-a-row cadence, not a revise-the-prose cadence. Distinct from reference
docs.§Ledger Anchoring section
regen-signingshipped 2026-04-17, produced 4 mainnet anchors onregen-1(Apr 17/18). Anchoring a KOI object produces a canonical/manifest/receipt triad viaMsgAnchor. v1.1.0 was silent on this entire discipline. The new section also documents rename discipline — anchored files are rename-hostile because modifying the file modifies its content hash, silently invalidating the on-chain anchor. A manifesto-level classification procedure (anchored-historical vs. forward-reference vs. anchored-living) prevents this failure mode.§Legacy Names
On-chain anchors predating v1.2.0 ratification use first fields that aren't canonical relevance values (e.g.,
strategy.plan.claims-engine-dogfood-plan.v0.1.0wherestrategyis a type, not a relevance). These cannot be retroactively corrected without invalidating the anchor. The section formalizes grandfathering — legacy names preserved as historical record; all new anchored objects use canonical[relevance].[type].[subject].vX.Y.Zgoing forward.Pilot Evidence
CONTRIBUTING.md requires a minimum of 3 objects using the proposed convention, piloted for at least 2 weeks. This PR includes 4 RegenOS Phase 0 artifacts as the pilot set — one per new type — all produced today (2026-04-21) as part of the RegenOS unified-operating-system scaffolding session:
core.spec.regen-os-learning-subroutines.v0.1.0specexamples/core.plan.regen-os-phase-delivery.v0.1.0planexamples/core.overview.regen-os-current-state.v0.1.0overviewexamples/core.registry.regen-os-artifact-registry.v0.1.0registryexamples/Plus an additional canonical entry already in the KOI Repo:
core.strategy.regen-os-overview.v0.1.0(valid under v1.1.0; companion to the four v1.2.0 pilots).Pilot duration note: this is day zero of piloting. Proposing the PR for review now, with the governance call as the ratification forum. Pilot can continue during the review window to satisfy the 2-week minimum before ratification.
Bugfixes
docs/semantic-naming-properties.md— the worked examplecore.objective.rnd-foundation-sprint.v2025-Q2.v0.1.4(appearing in three places: Notion property table, Google Docs metadata block, GitHub YAML frontmatter) had three issues:objectiveis not a defined type in any version of the manifestov2025-Q2.v0.1.4is a double-version — the versioning scheme is single-semverStatus: activeis not in the canonical status list (draft/in review/ready to publish/published)Replaced all three occurrences with a canonical v1.2.0 example:
core.strategy.regen-os-overview.v0.1.0.CONTRIBUTING.md— the "Who Can Contribute?" section header was missing its##prefix (line 6), breaking Markdown rendering. Restored.Scope Check
Version bump is minor (v1.1.0 → v1.2.0):
Test Plan
core.objective,v2025-Q2,Status: active)🤖 Generated with Claude Code