Persistent memory, BM25 code search, MCP tools, and hook guardrails for AI coding agents — Claude Code, Codex CLI & Antigravity. One local-first .ai/ layer per repo. No SaaS, no network on hot paths.
This repository also configures its own agents. See what code-brain tells them →
Agent Claude Code
A separate agent for reviewing complex codebase decisions, large refactors, and difficult cause analysis. A refactor changes internal code structure while aiming to preserve behaviour.
Agent Claude Code
A separate agent for researching an existing codebase, including its patterns, affected areas, and similar implementations. It reports evidence from specific files, functions, and lines without changing files.
Agent Claude Code
A separate agent that reviews security-sensitive code changes without editing files or reading sensitive configuration. It focuses on authentication, permissions, payments, deletion, and handling of outside input.
Agent Claude Code
A test-reviewing subagent that checks whether code changes have enough tests and identifies how they should be verified.
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: