Motir

⚠️ CORRECTED 2026-08-06 at the start of its own run — THREE of the four gaps this card claimed DO NOT EXIST. The card was written from a sweep that listed only person-NAMED keys (userId, authorId, …); name was not in the pattern, so a nested { userId, name } looked like { userId }. Reading the schemas settled it. Same shape as notes.html #229 — a conclusion drawn from a tool's output without opening the file — caught this time before any code was written.

Verified against lib/api/v1/workLoop/schema.ts and workItems/schema.ts:

The activity CHANGE side already got this right. Only the COMMENT side applied the rationale Amendment 9 Q1 overturned — and presentComment's own comment says so verbatim: "a public API must not acquire a second, accidental user resource."

So this card is ONE widening, not four.

What to build

commentSchema gains author: { id, name } beside authorId. That is the entire change. It reaches both consumers at once, because commentThreadSchema extends commentSchema and activityEntrySchema's comment variant reuses it verbatim.

• actorRefSchema MOVES from lib/api/v1/ready/schema.ts to lib/api/v1/workItems/schema.ts, and ready/schema.ts imports it from there. Required, not cosmetic: ready/schema.ts already imports FROM workItems/schema.ts, so declaring the shared shape in ready would invert the dependency.
• CommentSource.author widens from { id: string } to { id: string; name: string } — the narrowing that caused this lives in that interface, not in the service. CommentAuthorDTO already carries name (lib/dto/comments.ts:8), so no new query: it is a mapper widening exactly like 11.5.13's.
• presentComment's comment is rewritten — it currently states the overturned rationale as fact. A schema whose prose contradicts the ADR is the defect 11.5.13's own criteria named.
• Keep authorId. Removing it is a §8 violation.
• V1_CONTRACT_VERSION → 1.3.0 (1.1.0 was MOTIR-2275's header, 1.2.0 was 11.5.13's ready-row actor).

Scope BOUNDARY

Ends at commentSchema + the actorRefSchema move. ONE repo, motir-core. It does NOT change the CLI (11.5.4 consumes it), does NOT touch the change/value schemas (verified correct above), and does NOT add a user endpoint.

Acceptance criteria

• commentSchema carries author: { id, name } beside authorId, and a test asserts author.id === authorId.
• actorRefSchema is declared once, in workItems/schema.ts, and ready/schema.ts imports it — no duplicate declaration.
• render.ts's renderComment / renderCommentThread are driven from a REAL GET /api/v1/work-items/{key}/comments and a REAL …/activity?view=all response and print the author's NAME — added to tests/api/v1/cli-renderers-from-v1.test.ts.
• No new service call: asserted by spying on the comments read over a multi-comment page.
• The field appears in the emitted document; V1_CONTRACT_VERSION moves to 1.3.0 in the same PR.
• 11.6's MCP drift guard still passes.
• presentComment's prose agrees with Amendment 9 afterwards.
• The per-file coverage floor (≥90%) holds.

Context refs

• docs/decisions/public-api-conventions.md Amendment 9 Q1 — the rule this applies.
• lib/api/v1/workItems/schema.ts — commentSchema, CommentSource, presentComment, and actorRefSchema's new home.
• lib/dto/comments.ts:8 — CommentAuthorDTO, which already carries the name.
• lib/api/v1/workLoop/schema.ts:401 · :452 — the two shapes that were ALREADY right; do not touch.
• packages/cli/src/render.ts:915 — the reader.
• Blocks: 11.5.4.