harness

Agent engineering harness for any repo. Creates a short AGENTS.md table-of-contents, structured docs/ knowledge base (ARCHITECTURE, QUALITY, CONVENTIONS, COORDINATION), custom agent-readable linters (WHAT/FIX/REF format), CI enforcement, and execution plan templates. Supports Rust, Go, TypeScript, and Python. Use when setting up any repo for agent-first development, upgrading an existing AGENTS.md, or enforcing architectural lint gates. Includes --audit flag for tool lifecycle checks and L1/L2/L3 progressive disclosure.