What problem does it solve? Personal knowledge bases accumulate sensitive material — compensation details, performance assessments, retention conversations, and candid opinions — that must be removed before the brain can be shared with a team. Clean working files alone are insufficient because git history and structured takes/facts still carry the sensitive data. ## Core Features & Use Cases - Full-scope sanitization: Scans people, meetings, dailies, companies, projects, and analysis directories plus takes and facts, removing ratings, compensation, performance assessments, and political dynamics while preserving professional identity and work history. - Gated destructive operations: Fires data-loss-gate confirmation cards before bulk content stripping and before git history rewriting with force-push, requiring an explicit typed yes. - Verification and re-audit: Runs grep triage plus retrieval-based checks (gbrain query, takes search, recall) against the tree that ships, and supports report-only periodic re-audits of an existing shared brain. - Use Case: A founder wants to stand up a shared company brain from their personal brain. The skill copies the scanned directories to a staging tree, strips sensitive content, verifies with retrieval queries, and exports a fresh sanitized repo for the team. ## Quick Start Ask the agent to sanitize my personal brain into a team-safe shared brain, stripping compensation, ratings, and performance notes before anyone else gets access.
