Skip to main content
Glama
deenesjakoruzh
jezdez

conda-cli-mcp

by jezdez
OverviewSchemaSourceRelated ServersDiscussionsAdmin
Python
Local

conda-cli-mcp

conda-cli-mcp exposes an installed conda command line through the Model Context Protocol. It discovers built-in and plugin-contributed commands at startup, creates structured tools from conda's argparse definitions, and keeps the complete CLI reachable through an explicitly enabled raw argv tool.

The server runs conda in bounded child processes. Conda remains responsible for configuration, plugin ordering, solving, transactions, reporters, authentication, hooks, and error handling.

When conda-completion 0.3 or newer is installed in the target environment, a read-only metadata tool exposes its command-specific completion hints through the plugin's public Python API.

The project is alpha software. It requires Python 3.10 or newer and supports target conda versions >=26.7,<27.

Run the standalone server as ccm. When the package is installed in conda's own environment, conda mcp launches the server against that conda installation without requiring --conda-exe. These launchers require conda-cli-mcp 0.2.0 or newer.

Installation

Install from the jezdez channel on Anaconda.org:

conda create --yes --name conda-cli-mcp --override-channels --channel jezdez --channel conda-forge "conda-cli-mcp>=0.2"
conda activate conda-cli-mcp
ccm --version

Related MCP server: terminal-guardian-mcp

Documentation

The documentation is organized by purpose:

Documentation is published at jezdez.github.io/conda-cli-mcp.

Safety

The default policy is read-only. --allow-write permits conda-managed mutations. --allow-exec permits conda run, opaque passthrough arguments, and the raw conda_execute tool. Policy is fixed at startup and enforced before conda starts.

Running the server

Select the target conda executable explicitly when it lives in another environment:

ccm --conda-exe /absolute/path/to/conda

Without --conda-exe, resolution checks CONDA_EXE and then PATH. MCP clients connect over standard input and standard output. See the getting-started tutorial for a client configuration.

Project scope

  • conda-cli-mcp owns local conda CLI discovery and execution.

  • conda-meta-mcp owns read-only package and ecosystem metadata.

  • conda-runtime owns the ready-to-run conda executable and can include this plugin as a locked runtime dependency.

  • conda-mcp is reserved for a future umbrella package that installs both.

Development

Development, documentation, and packaging use Pixi:

pixi install
pixi run test
pixi run check
pixi run -e docs docs
pixi run -e build package

The conda recipe is noarch Python and lives in recipe/recipe.yaml.

This server cannot be deployed

One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

Maintenance

ActivityMaintained
ResponsivenessNo issues

Resources

Related MCP Connectors

Related MCP Servers

  • AWS MCP Server

    saifeezibrahim
    A
    license
    Not graded
    quality
    D
    maintenance
    Enables AI assistants to execute AWS CLI commands and access AWS documentation, resources, and prompt templates through the Model Context Protocol with support for Unix pipes and secure Docker-based deployment.
    MIT

  • terminal-guardian-mcp

    7Majesty-M
    A
    license
    Not graded
    quality
    D
    maintenance
    Provides secure, sandboxed terminal access for AI assistants via the Model Context Protocol, with multi-layer risk analysis and auditable command execution.
    13 npm
    8
    MIT