drhalto/agentmako

What it does

agentmako is a codebase intelligence engine designed to enhance AI coding agents with deep repository understanding. It provides structured context packets, indexed code and schema facts, and diagnostic information through the Model Context Protocol (MCP). Rather than having agents repeatedly scan repositories with grep or similar tools, agentmako maintains a local SQLite-backed knowledge store of your codebase, enabling agents to query relevant code context efficiently.

How to set up

agentmako operates as an MCP server that can be integrated into compatible AI platforms. Install the server from the drhalto/agentmako GitHub repository and configure it with your local codebase path. The tool uses a local SQLite database by default but supports optional Postgres or Supabase integration for enhanced collaboration. Once configured, connect your AI agent to the MCP server endpoint to begin querying structured codebase facts.

Key features

• Local-first SQLite-backed codebase indexing
• Structured context packet generation for AI agents
• Indexed code and schema facts for efficient querying
• Diagnostic information and repository insights
• Optional Postgres/Supabase support
• MCP protocol integration for broad AI platform compatibility
• Eliminates redundant repository scanning by agents