Resolves #13618
Aligns the remaining live skill payloads with §no_hold_state: stale wake/context-recovery/pre-review guidance now routes gated or absent current lanes into another named lane instead of preserving retired terminal vocabulary. This completes the AC4 prose tail after the earlier validator and AGENTS changes; #13616 already carries the public closure/redirect correction for the discussion recurrence.
Evidence: L2 static substrate validation achieved; L3 runtime proof is not required because this is documentation-only agent-governance payload cleanup. Residual: none for #13618.
Deltas from ticket
Turn-Memory Pre-Flight / Slot Rationale
- Placement: conditional skill payloads only, not always-loaded
AGENTS.md or SKILL.md routers.
- Disposition delta:
rewrite stale terminal wording into existing §no_hold_state references.
- Trigger-frequency: only when peer-role, context-recovery, or pre-review intake payloads are explicitly loaded.
- Failure-severity: high for no-hold regressions, but the correct fix is payload alignment because Layer 1 and the stop hook already carry the always-loaded authority.
- Enforceability: discipline-only prose alignment; no new hook, validator state, or rule layer added.
Test Evidence
rg -n "verified-empty|human-gate|blocked-task-state" .agents/skills/peer-role/references/peer-role-mode.md .agents/skills/context-recovery/references/context-recovery-workflow.md .agents/skills/post-review-pickup/references/pre-review-intake-lane-gate.md returned no matches.
node ai/scripts/lint/lint-skill-manifest.mjs --base origin/dev passed.
git diff --check passed.
npm run agent-preflight -- .agents/skills/peer-role/references/peer-role-mode.md .agents/skills/context-recovery/references/context-recovery-workflow.md .agents/skills/post-review-pickup/references/pre-review-intake-lane-gate.md passed.
- Commit hook ran
check-whitespace.mjs successfully.
Post-Merge Validation
Commit
e157bcd617 — fix(agentos): align no-hold skill prose (#13618)
Authored by Euclid (GPT-5, Codex Desktop). Session 1c4b42c3-289a-4196-bec0-36a3a9f16fa6.
Resolves #13618
Aligns the remaining live skill payloads with
§no_hold_state: stale wake/context-recovery/pre-review guidance now routes gated or absent current lanes into another named lane instead of preserving retired terminal vocabulary. This completes the AC4 prose tail after the earlier validator and AGENTS changes; #13616 already carries the public closure/redirect correction for the discussion recurrence.Evidence: L2 static substrate validation achieved; L3 runtime proof is not required because this is documentation-only agent-governance payload cleanup. Residual: none for #13618.
Deltas from ticket
Turn-Memory Pre-Flight / Slot Rationale
AGENTS.mdorSKILL.mdrouters.rewritestale terminal wording into existing§no_hold_statereferences.Test Evidence
rg -n "verified-empty|human-gate|blocked-task-state" .agents/skills/peer-role/references/peer-role-mode.md .agents/skills/context-recovery/references/context-recovery-workflow.md .agents/skills/post-review-pickup/references/pre-review-intake-lane-gate.mdreturned no matches.node ai/scripts/lint/lint-skill-manifest.mjs --base origin/devpassed.git diff --checkpassed.npm run agent-preflight -- .agents/skills/peer-role/references/peer-role-mode.md .agents/skills/context-recovery/references/context-recovery-workflow.md .agents/skills/post-review-pickup/references/pre-review-intake-lane-gate.mdpassed.check-whitespace.mjssuccessfully.Post-Merge Validation
verified-empty,human-gate, orblocked-task-stateas terminal exits.Commit
e157bcd617—fix(agentos): align no-hold skill prose (#13618)Authored by Euclid (GPT-5, Codex Desktop). Session 1c4b42c3-289a-4196-bec0-36a3a9f16fa6.