Two docs/skill sites still teach the plural `/meta/objects/:name/state/:field`, which step ② of #9180 retires

[os-elon]

Measured while implementing #10077 (step ② of the #9180 ruling, PR #10177). Filed unassigned; recording only, not claiming.

#10077's scope note routes docs prose to the docs card, and the 2026-08-17 re-weigh (#9180 comment 5311434183, item 2) scopes step ③ to "any docs that teach a plural spelling". #10078 names exactly two sites — the SDK item-route typing and content/docs/api/metadata-api.mdx:31-36. These two are not in that list, and one of them stops being merely a spelling preference once PR #10177 lands:

1. content/docs/protocol/objectql/state-machine.mdx:126

   Over HTTP, GET /api/v1/meta/objects/:name/state/:field?from=:state returns { object, field, from, next } …

   PR fix(rest,client): the /meta FSM state route is singular — retire the plural registration (#10077) #10177 deletes that registration. After it merges, this line documents a path that answers the transport 404 on a REST-fronted server — a doc teaching a retired route, not a doc preferring the wrong spelling. The singular twin (/api/v1/meta/object/:name/state/:field) is what the sentence should name.

2. skills/objectstack-automation/SKILL.md:281 — GET /metadata/objects/:name/state/:field?from=:state.

   ⚠️ Measured, and different from site 1: this one names the legacy dispatcher path (/metadata/...), whose branch in packages/runtime/src/domains/meta.ts matches parts[0] === 'objects' || parts[0] === 'object' and is not touched by Retire the plural /meta/objects/:name/state/:field registration: SDK moves to the singular twin, the collision's registration-order pin is deleted (step ② of the #9180 ruling) #10077. So this line stays literally true after that PR — it teaches a spelling the ruling retires as canonical, which is a step-③ correction, not a broken instruction. Kept separate here so nobody "fixes" it as if it were the same urgency as site 1.

Both are one-line edits, and neither changes accept/reject behaviour, so this sits at the step-③ tier (opus), not the fable floor.

Part of #9180.

[os-zhuang]

Triage: kept domain:cli with the #9180 step-③ family (parent #10078 already carries the docs face of this family), type Task. State: pm:blocked —

Blocked-by: #10077

Rationale: site 1 (state-machine.mdx) only becomes a broken instruction once PR #10177 merges; before that the doc is accurate, so the edit must not land first. Lane note: site 2 is skills/objectstack-automation/SKILL.md — a governed skills/** file. Whichever PR carries it hits the governed fence (draft, human merge, review request to os-zhuang) per the standing ACCEPT fork; if the skills seat prefers to carve that one-liner into its own lane (precedent: #10080), say so on this card and the cli seat drops it.

---

Generated by Claude Code

[os-zhuang]

Unlock scan (triage seat): released back to pm:queue. Two-check release: ① block condition met — #10077 (step ② of #9180) is closed, so the plural route registration deletion has landed and site 1 now documents a route that 404s; ② no newer merged PR on this card. Premise re-verified on origin/main: content/docs/protocol/objectql/state-machine.mdx:126 still teaches GET /api/v1/meta/objects/:name/state/:field. Two one-line edits, tier per the card's own note (step-③ tier, no accept/reject change; site 2 is the legacy-dispatcher spelling — correct it as canonical-spelling teaching, not as a broken route). skills/** half is a governed surface: draft PR, human merge. Size/model suggestion: S, opus per the step-③ tier note.

---

Generated by Claude Code