# AI Policy Check

> Check compliance with AI_POLICY.md when creating a git commit or a pull request in this repository.

- Skill: `vivliostyle/ai-policy-check-2` (Agent Skill)
- Install (CLI): `npx get-skill add vivliostyle/ai-policy-check-2`
- Raw SKILL.md: https://api.skillmd.com/api/skills/vivliostyle/ai-policy-check-2/raw
- Safety review: pending (external: skill-scanner PASS, skillspector PASS)
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: vivliostyle (https://skillmd.com/u/vivliostyle)
- Updated: 2026-09-22
- Page: https://skillmd.com/skills/vivliostyle/ai-policy-check-2

---


# AI Policy compliance check

For commits and pull requests that contain AI-generated changes (see
`AI_POLICY.md` at the repository root):

<!--
  The "does not need to be last" / Co-authored-by clauses below were added
  after a Copilot review on PR #57 found this checklist stricter than
  AI_POLICY.md itself, which explicitly accepts tool-added Co-authored-by:
  trailers as disclosure. Keep this checklist in sync with the policy text
  rather than re-tightening it independently.
-->
- **Commit**: the message includes an `Assisted-by: AGENT_NAME:MODEL_VERSION`
  trailer, naming the exact model used, e.g.
  `Assisted-by: Claude Code:claude-sonnet-5`. It does not need to be the last
  trailer. Tool-added `Co-authored-by:` trailers (e.g. from GitHub Copilot on
  PR review commits) are also accepted as disclosure, per AI_POLICY.md.
- **PR body**: contains the same `Assisted-by:` line, and a short description
  of the human/AI division of labor inside a
  `<details><summary>How the AI was used</summary>` block (see the example in
  `AI_POLICY.md`).
- **Content**: technical claims are verified against the actual source
  (running code or `submodules/`), not written from assumption. No
  third-party copyrighted material is reproduced.

Fix anything that does not match before pushing or opening the PR.

