# Product Playbook Audit

> Audit an existing product playbook against current evidence without editing Markdown or state. Use when the user asks for a playbook audit, drift check, coverage gap review, or read-only reconciliation report. Prefer the main product-playbook skill for create or edit runs.

- Skill: `b-amir/product-playbook-audit` (Agent Skill)
- Install (CLI): `npx skillmds@latest add b-amir/product-playbook-audit`
- Raw SKILL.md: https://api.skillmd.com/api/skills/b-amir/product-playbook-audit/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Security
- Author: b-amir (https://skillmd.com/u/b-amir)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/b-amir/product-playbook-audit

---


# Product Playbook Audit

Read-only companion to `product-playbook`. Same discovery and evidence rules.
Never write playbook Markdown, state, PDF/HTML, or findings unless the user
explicitly switches mode.

## Steps

1. Run bootstrap / inventory with `--check-state` (add `--drift` for CI exit codes).
2. Run bootstrap and print `intake.intake_message` verbatim (table + Correct me if I'm wrong +
   lettered choices with blank lines preserved). No polls for Intake. Same sources → same message.
3. Report Keep / Update / Add / Remove / Needs evidence — do not patch.
4. Emit End Report with `Mode: audit`.

Skill root scripts live two levels up: `../../scripts/`.

```bash
python3 <skill-dir>/../../scripts/inventory_playbook.py "<output_dir>" \
  --source "SOURCE_ID=<root>" \
  --check-state --drift
```

Full protocol: [../../references/run-protocol.md](../../references/run-protocol.md).