Fix the audit-docs fixing documentation out of scope

[AlexJSully]

No description provided.

[vercel]

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

Project	Deployment	Actions	Updated (UTC)
portfolio	Ready	Preview	Aug 7, 2026 6:46pm

[Copilot]

Pull request overview

This PR updates the repository’s agent prompt and skill documentation to prevent audit-docs from “fixing” documentation and comments outside the scope it resolved (especially on pull requests), and adds guardrails so bundled procedure outputs are treated as leads to verify rather than publishable sources. It also includes a small dependency upgrade.

Changes:

• Tightens audit-docs scoping language so Phase 2 and Phase 3 operate only on the documents and code the scope rule resolved to, not an automatic full docs/ sweep.
• Reframes “subagents” as “procedures” and adds explicit guidance that procedure/agent outputs are not publishable evidence without re-grounding.
• Adds optional skill plugin manifests and extends the publishability checker to validate plugin manifests and ensure published skills remain invocable by name.

Reviewed changes

Copilot reviewed 16 out of 17 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
package.json	Bumps Sentry and Sass versions.
package-lock.json	Lockfile updates for Sentry/Sass and transitive dependency changes.
CLAUDE.md	Clarifies the skill invocation name format for /typescript-code-and-test-standards.
.github/prompts/readme.md	Minor wording updates describing typescript-code-and-test-standards behaviour.
.github/prompts/audit-pr.prompt.md	Strengthens refutation-pass guidance about re-grounding verdicts.
.github/prompts/audit-docs.prompt.md	Narrows audit scope language for docs and in-code comment phases.
.github/copilot-instructions.md	Tightens guidance around deleting duplicated comments only when safely verified.
.claude/skills/typescript-code-and-test-standards/SKILL.md	Removes paths: from published skill frontmatter and reframes “subagents” as bundled procedures.
.claude/skills/typescript-code-and-test-standards/.claude-plugin/plugin.json	Adds plugin manifest metadata for the skill.
.claude/skills/audit-pr/SKILL.md	Adds guidance on running finding-refuter.md as a procedure and re-grounding its verdicts.
.claude/skills/audit-pr/.claude-plugin/plugin.json	Adds plugin manifest metadata for the skill.
.claude/skills/audit-docs/SKILL.md	Updates bundled-procedure guidance and narrows doc/comment audit scoping language.
.claude/skills/audit-docs/agents/surface-auditor.md	Requires explicit path lists and adds bounds to avoid drifting outside scope.
.claude/skills/audit-docs/agents/coverage-mapper.md	Requires explicit document/code path lists and rejects name-based scope.
.claude/skills/audit-docs/.claude-plugin/plugin.json	Adds plugin manifest metadata for the skill.
.claude/scripts/check-skill-publishability.mjs	Validates optional plugin manifests and enforces “published skills stay invocable by name” policy.
.claude/rules/prompt-skill-sync.md	Documents plugin manifest behaviour/limitations and published-skill invocability policy.