Persistent code-knowledge memory for AI chats. An MCP server that caches a codebase's structure and summaries in a local SQLite "brain" so sessions query it instead of re-reading files — token-cheap, 100% local.
marcodavidd020/cerebro-code-memory
Agent
Use BEFORE writing code for a feature or a new project — this is the design gate that prevents hallucinated work. It (1) grounds itself in the project's REAL standards (CLAUDE.md + Cerebro), (2) asks you in detail what functionality to include/design, and (3) returns a grounded spec/plan to approve. It does NOT write…
marcodavidd020/cerebro-code-memory
Agent
Use to ANSWER questions about an existing codebase — "where is X?", "how does Y work?", "what depends on Z?", "explain this module or flow" — WITHOUT spending tokens re-reading folders. Read-only navigator that queries the Cerebro brain first and only opens files when the brain has no/stale info. Use PROACTIVELY…
marcodavidd020/cerebro-code-memory
Agent
Use BEFORE a refactor or risky change to assess blast radius and architectural health, or to audit dead code and circular dependencies. Read-only. Answers "what breaks if I change X?", "is this safe to delete?", "where are the import cycles?", "which modules are most fragile?". Triggers (es): "qué riesgo tiene…
marcodavidd020/cerebro-code-memory
Agent
Use to IMPLEMENT a focused code change — fix a bug, add a small feature, refactor a function — once the goal is clear. Rebuilds context cheaply via Cerebro before editing, makes surgical changes that match surrounding code, verifies, then records what it learned and decided back into the brain. Triggers (es): "arregla…
marcodavidd020/cerebro-code-memory
Agent
Use to keep the Cerebro brain fresh and well-documented so future sessions stay cheap. Run after a work session, after a git pull / branch switch, or periodically: it syncs and reindexes changed files, finds important files lacking summaries, and records concise summaries plus loose decisions. Triggers (es)…