extract-rules-from-vault

To extract tribal knowledge from a Notion or Slack export, this Skill parses deterministic signals like channels, headings, and recurring phrases before synthesis. It outputs draft hookify rules and a CLAUDE.md file for immediate founder review.

Populating a new Claude install with company documentation involves walking a folder of markdown or GDocs artifacts to emit draft skills and memory files. This approach ensures your fresh install starts with encoded rules rather than an empty state.

Yes, you can generate draft hookify rules directly from a folder of markdown vault files. The extractor operates in a structured-signal-first mode, deriving deterministic signals from file paths and headings before synthesizing reusable drafts.

Converting document exports into structured signals and draft skills requires parsing deterministic artifacts like users and paths first. The model then performs synthesis on residual data to produce candidate skills and a signals.json file for review.

When extracting patterns from company artifacts for AI onboarding, outputs include a CLAUDE.md.draft, hookify-rules, candidate skills, signals.json, and an extraction-report. These files include a REVIEW.md gate for final acceptance before live deployment.

Yes, you need to review draft CLAUDE.md files generated from Slack exports before deployment. The outputs are specifically designed for founder review and refinement, utilizing a REVIEW.md gate to ensure formal live acceptance of the extracted rules.