Install the content verification protocol (no fabricated facts, quotes, numbers or visuals)
Ready nowFour rules that stop fabricated facts: auto-transcripts lie, drafts are not facts, prospects are not customers, unusual names get verified; quotes only from a downloaded transcript; mock numbers never quoted as real; every visual rendered to PNG and read before it is called done.
Kit
No separate kit. The playbook is self-contained.
You need
| Account | Why | Free or paid | Plan |
|---|---|---|---|
| yt-dlp | Downloads YouTube, TikTok and Instagram video, auto-captions and comments for vaults, B-roll, MIDI extraction and quote verification. | Free | Free. |
| Playwright + headless Chrome (and Safari for the final check) | Renders pages to PNG for visual QA, drives the dashboard verification protocol, captures library stills, renders HTML titles and animations offline. | Free | Free. Google Chrome installed for headless screenshots; Playwright installs its own Chromium and WebKit. |
| Claude Code (Anthropic Claude Pro or Max) | The AI operator every system runs through. Skills, hooks, memory, MCP servers and the Brain plugin all load inside Claude Code. | Paid | Claude Pro (20 USD/mo) works for light use; Max (100 USD or 200 USD/mo) for daily multi-hour sessions. Anthropic terms require each user to hold their own subscription; never share a login. |
What you bring
- Your press kit or LinkedIn for every person you name
The steps
- Put the four rules and the pre-flight checklist in CLAUDE.md
- Build the verified-facts and unverified-flags files
- Gate behavioral claims about real people
- Verify every quote against the transcript, never against a search summary
- Quarantine mock data
- Render and look at every visual before "done"
- Audit the state before prescribing a fix
- Cold-read before ship
Goes further if you have already built Turn every correction into a memory your AI keeps. You can build this without it.
Get the full playbook
Part of: Get your database sorted