Context
Enforcement deliverable of course lane 8 (#2910; rubric locked 2026-08-17, register gate clean, user-confirmed). The rubric is docs/conventions/invocation-mode/README.md (convention-registry row in docs/PLUGIN-PHILOSOPHY.md) — this issue executes its enforcement; do not relitigate the locked decisions. Decision rows: docs/upstream/aihero-steering-lanes.md (lane 8). Contract: docs/specs/invocation-mode-doctrine-brief.md.
Fleet count note: the lane graded against the 2026-08-17 measurement (211 skills); the chain-close merge brought the fleet to 215 (four incoming skills all explicitly false, default-conformant — recount recorded in the rubric doc). The criterion below is count-agnostic.
Proposed work
- Add explicit
disable-model-invocation: false to the 17 skills currently missing the key (all sit in the rubric's default class; the official default for an absent key is already false, so this is a no-behavior-change normalization): code-tidying:batch-simplify, education:explain, education:quiz-me, event-storming:methodology, event-storming:simulation, github:advise, github:audit, knowledge:book-distill, machine-health:audit, naming:name-it-better, plugin-quality:audit, repo-fleet-hygiene:audit, repo-hygiene:clean, toolchain:check, toolchain:lint, visualization:visualize, wizard:generate
- New
skill-quality:check criterion: every skill's frontmatter carries an explicit disable-model-invocation key; a true value must be attributable to one of the rubric's three exception classes
playbooks:skill-authoring: authoring-time pointer to the rubric (point, don't restate)
- Per-plugin version bumps + CHANGELOG entries for every touched plugin
Acceptance criteria
Metadata
| Field |
Value |
| Category |
unspecified |
| Area |
skill-quality, playbooks, all plugins (fleet), docs (conventions) |
| Ecosystem |
unspecified |
Context
Enforcement deliverable of course lane 8 (#2910; rubric locked 2026-08-17, register gate clean, user-confirmed). The rubric is
docs/conventions/invocation-mode/README.md(convention-registry row indocs/PLUGIN-PHILOSOPHY.md) — this issue executes its enforcement; do not relitigate the locked decisions. Decision rows:docs/upstream/aihero-steering-lanes.md(lane 8). Contract:docs/specs/invocation-mode-doctrine-brief.md.Fleet count note: the lane graded against the 2026-08-17 measurement (211 skills); the chain-close merge brought the fleet to 215 (four incoming skills all explicitly
false, default-conformant — recount recorded in the rubric doc). The criterion below is count-agnostic.Proposed work
disable-model-invocation: falseto the 17 skills currently missing the key (all sit in the rubric's default class; the official default for an absent key is alreadyfalse, so this is a no-behavior-change normalization):code-tidying:batch-simplify,education:explain,education:quiz-me,event-storming:methodology,event-storming:simulation,github:advise,github:audit,knowledge:book-distill,machine-health:audit,naming:name-it-better,plugin-quality:audit,repo-fleet-hygiene:audit,repo-hygiene:clean,toolchain:check,toolchain:lint,visualization:visualize,wizard:generateskill-quality:checkcriterion: every skill's frontmatter carries an explicitdisable-model-invocationkey; atruevalue must be attributable to one of the rubric's three exception classesplaybooks:skill-authoring: authoring-time pointer to the rubric (point, don't restate)Acceptance criteria
disable-model-invocationkey (skill-quality:checkpasses fleet-wide with the new criterion)skill-qualityand cross-linked to the rubric docplaybooks:skill-authoringpoints atdocs/conventions/invocation-mode/README.mdMetadata