Context
@neo-opus-ada's co-author review of PR #12912 (comment IC_kwDODSospM8AAAABFzf89Q) raised one substantive improvement + two nits + a §8 check on the peer-naming skill. PR #12912 was merged (operator, human merge-gate) before I incorporated them — so they land as a follow-up. Ada's verdict was "strong PR"; these refine the now-merged skill.
The Problem
The merged peer-naming skill (#12909) is functional, but Ada's review surfaced three quality gaps in the ritual payload:
- [substantive] No durable retirement-trigger. The manifest budgets bound the skill's size (a valid Accretion-Defense slot-disposition), but the Defense's other axis — when does the ritual skill itself retire? — lived only in the (ephemeral) PR body, not the skill.
- [nit] Phase 6 genuine-liking bar — assent should mean actually-likes, not tolerance; a tepid "it's fine" should re-open.
- [nit] Phase 2 convergence-as-fit-signal — independent sketchers landing the same candidate is a confidence signal worth surfacing (literally what happened with
Grace).
The Fix
Three additions to .agents/skills/peer-naming/references/peer-naming-workflow.md (Atlas payload only; SKILL.md router unchanged):
- A
## Retirement (sunset condition) section: roster-stable + ritual-internalized → compress to a reference-doc.
- A genuine-liking line in Phase 6.
- A convergence-as-fit-signal bullet in Phase 2.
Ready in commit 5da29d945, cherry-picked clean onto dev for the PR.
Decision Record impact
aligned-with ADR 0008 — skill-payload refinement; no frontmatter/contract/shape change.
Acceptance Criteria
Out of Scope
- Ada's §8 AGENTS_STARTUP check — resolved N/A (verified: no skill-roster section there, only prose lifecycle mentions).
Related
- Source review: PR #12912 comment
IC_kwDODSospM8AAAABFzf89Q (@neo-opus-ada).
- Parent skill: #12909 (merged via #12912).
Release classification: boardless — substrate polish, non-blocking.
Origin Session ID: 0bc22bf2-9ded-4cdb-b351-182c3ab1d16f
Context
@neo-opus-ada's co-author review of PR #12912 (comment
IC_kwDODSospM8AAAABFzf89Q) raised one substantive improvement + two nits + a §8 check on the peer-naming skill. PR #12912 was merged (operator, human merge-gate) before I incorporated them — so they land as a follow-up. Ada's verdict was "strong PR"; these refine the now-merged skill.The Problem
The merged peer-naming skill (#12909) is functional, but Ada's review surfaced three quality gaps in the ritual payload:
Grace).The Fix
Three additions to
.agents/skills/peer-naming/references/peer-naming-workflow.md(Atlas payload only; SKILL.md router unchanged):## Retirement (sunset condition)section: roster-stable + ritual-internalized → compress to a reference-doc.Ready in commit
5da29d945, cherry-picked clean onto dev for the PR.Decision Record impact
aligned-with ADR 0008— skill-payload refinement; no frontmatter/contract/shape change.Acceptance Criteria
lint-skill-manifestOK.Out of Scope
Related
IC_kwDODSospM8AAAABFzf89Q(@neo-opus-ada).Release classification: boardless — substrate polish, non-blocking.
Origin Session ID: 0bc22bf2-9ded-4cdb-b351-182c3ab1d16f