Guard JudgmentKit deck skill discoverability

[mikeylong]

Summary

• add Codex plugin verification checks for JudgmentKit slide-deck workflow language in the hosted skill
• require OpenAI skill metadata to remain discoverable for slide deck / PPTX creation prompts
• add regression tests for stale skill text and stale agent metadata

Verification

• npm run codex-plugin:check
• node tests/codex-plugin.test.mjs
• npm run codex-plugin:build
• git diff --check

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
judgmentkit-ai	Ready	Preview, Comment	Jul 7, 2026 7:02pm

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: dcff51a7cc

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".

[chatgpt-codex-connector]

Require deck terms in skill frontmatter

For Codex skill selection, the discoverable skill summary comes from the SKILL.md frontmatter description, but this new happy-path fixture adds deck/PPTX wording only in the body while leaving the fixture's description: UI-only. Since collectMissingWordingFailures() searches the whole file, deleting the slide-deck terms from the real frontmatter would still pass codex-plugin:check as long as these body snippets remained, so the new guard does not catch the discoverability regression it is meant to prevent. Please parse and check the frontmatter description, or include a fixture that fails when only that metadata goes stale.

Useful? React with 👍 / 👎.