01Provides semantic search across months of Claude Code session history using LanceDB vector embeddings
02Monitors session health, tracks corrections, and generates handoff briefs for new sessions
03Enforces prompt discipline with 12 specialized tools to prevent ambiguous AI instructions
04Automates code verification, including type checks, tests, and build processes, before declaring tasks complete
050 GitHub stars
06Supports local (Xenova) or OpenAI-powered embeddings for efficient history indexing