https://github.com/ptonelli/mcp-python
a set of mcp tools to host on your own server with a shell and python interpreter
https://github.com/ptonelli/mcp-python
git mcp-server python3 shell
Last synced: 12 months ago
JSON representation
a set of mcp tools to host on your own server with a shell and python interpreter
- Host: GitHub
- URL: https://github.com/ptonelli/mcp-python
- Owner: ptonelli
- Created: 2025-06-23T07:53:59.000Z (about 1 year ago)
- Default Branch: master
- Last Pushed: 2025-07-07T20:29:20.000Z (12 months ago)
- Last Synced: 2025-07-07T23:24:40.469Z (12 months ago)
- Topics: git, mcp-server, python3, shell
- Language: Python
- Homepage:
- Size: 35.2 KB
- Stars: 0
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Rationale
I want my LLM to be able to read, modify and update my code. this implies it
should be able to:
- read and write files and directories
- run code (at least python)
- install dependencies (uv or venv)
Now on how I want to do it: This should run on its own : no need for an
additional machine or API access. The LLM must not have the ability to run
containers. The setup must itself be running inside a container with a mounting
point for data to easily run on a home server.
# Organisation
2 sets of tools
- shell prompt
- code execution
# Shell prompt (WIP)
Just provide a shell prompt with the ability to set the current active directory.
# Python execution (WIP)
No complex security, the python code must run and the LLM must be able to add its own deps to run the code.
# Integration with Librechat
To integrate MCP Python with Librechat, you need to update the following configuration files:
## docker-compose.yml
Add the following service configuration to your docker-compose.yml:
```yaml
mcp_python:
container_name: mcp_python
image: docker.nautil.org/mcp-python:latest
environment:
- WORKDIR=/home/projects
volumes:
- ./projects:/home/projects
```
## librechat.yaml
Add the following MCP server configuration to your librechat.yaml:
```yaml
mcpServers:
python:
type: streamable-http
url: http://mcp_python:8000/mcp
```
This setup allows Librechat to interact with the MCP Python service, providing code execution and file management capabilities to your LLM.