conda-meta-mcp
Verified Safeby conda-incubator
Overview
Provides read-only Conda ecosystem metadata to AI agents for answering packaging-related questions.
Installation
pixi run --manifest-path /path/to/pyproject.toml cmm runSecurity Notes
The project explicitly states a "Side-effect free by contract" design, with "No environment mutation" and "No external command side effects." Code review confirms adherence to this model, with no 'eval', 'exec', or direct file system writes for data storage beyond internal caching. It uses `requests.get` for `file_path_search` and `conda`'s internal mechanisms for `package_search` and `repoquery`, which involve legitimate network requests to fetch package metadata. All external I/O operations are wrapped in `asyncio.to_thread` for non-blocking execution, and `lru_cache` is used extensively to optimize performance and reduce repeated external calls. Input validation is present for many tools (e.g., `import_mapping`, `pypi_to_conda`, `file_path_search`) to prevent malformed queries.
Similar Servers
awesome-mcp-servers
A central directory for discovering and referencing various Model Context Protocol (MCP) servers, clients, and frameworks for AI agents.
mcp-forge
A command-line interface (CLI) tool for scaffolding new Model Context Protocol (MCP) servers with various configurable features and transports.
phone-a-friend-mcp
Enables real-time communication between multiple Claude Code instances, allowing one to listen and others to initiate conversations with specific intents.
qubinode_navigator
An AI-enhanced infrastructure automation platform that uses a multi-agent LLM architecture and RAG to manage hypervisor deployments, VMs, and cloud resources, providing context-aware guidance and troubleshooting.