KodHau

KodHau is an MCP server that injects team-specific engineering context into AI coding agents before they modify code. It analyzes closed GitHub pull requests to surface architecture decisions, rejected approaches, review comments, and undocumented constraints, then exposes that knowledge to Cursor, Claude Code, or any MCP-compatible client.

• PR-history analysis — extracts design decisions, constraints, and review context from closed pull requests • MCP integration — agents call kodhau_get_wisdom to retrieve relevant context for a file or function • Local-first workflow — product claims code and PR data stay on the user’s machine • Agent compatibility — works with Cursor, Claude Code, and other MCP clients • Team plans — scales from a free single-repo tier to team/business plans with shared knowledge and admin features

Useful for teams that want AI coding agents to respect historical architecture decisions instead of blindly editing code.