Broaden global skill consistency audits#133
Conversation
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 862e31992b
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| - Run `--mode skill --skill <skill-name> --repo-root <repo-root>` so common and | ||
| profile-specific source checks execute only for the selected skill. |
There was a problem hiding this comment.
When the global automation is the only routine audit, narrowing this step to --mode skill means each per-installed-skill invocation returns after check_selected_skills; the source validator's repo-wide checks for stale README rows, unknown $skill references in portable repo text, and other check_source_governance_consistency invariants only run on the full path. After a rename/removal or stale cross-repo reference outside the selected skill folder, no single-skill invocation covers the dropped repository-level deterministic contract even though the action still claims contract/doc/source consistency, so keep one full source validation per source repo or explicitly move those invariants into selected validation.
AGENTS.md reference: AGENTS.md:L81-L83
Useful? React with 👍 / 👎.
1 participant
Make skills-consistency-review target one manifest-backed installed skill regardless of name, add selected-skill runtime validation, and align the global automation contract and documentation. Checks passed: targeted runtime tests, full source validation, focused mypy, and Markdown lint.