orgbrain
Verified Safeby irab
Overview
Provides knowledge-extraction and cross-repository analysis for multi-repository organizations, enabling querying of type definitions, user flows, service dependencies, and infrastructure configurations across an entire codebase for AI agents.
Installation
MCP_CONFIG=config/repos.yaml node dist/index.jsEnvironment Variables
- MCP_CONFIG
Security Notes
The server uses `spawn('git', ...)` to interact with configured repositories, which involves cloning and potentially checking out code from URLs specified in `config/repos.yaml`. If a malicious repository URL were added to the configuration, this could pose a risk, such as execution of malicious git hooks or downloading harmful binaries. However, the system includes validation for repository names and file paths, uses `--bare` or `--depth 1` for initial clones (though `checkoutWorktree` creates a full worktree), and does not expose endpoints for arbitrary remote code execution via client requests. It's intended for use within a trusted development environment, with configuration managed by the user. No direct `eval` of untrusted inputs or hardcoded secrets were found.