Detect low-quality AI-generated code using a 3-lens parallel review architecture with model-tiered agents. Scans for AI authorship signals, idiom violations, and code quality issues.
Context window protection — sandboxed execution in 11 languages, FTS knowledge base with BM25 ranking, session continuity. Single Rust binary, <5ms hooks.
Central issue tracker for AI-assisted coding workflows. Manage tasks, track work, and maintain context with simple CLI commands.