Enterprise DNA
M MCP Servers Developer low

hileamlakB/PRIMS

by Various

PRIMS is a lightweight, open-source Model Context Protocol (MCP) server that lets LLM agents safely execute arbitrary Python code in a secure, throw-away sandbox.

H

MCP

hileamlakB/PRIMS

Added 1 June 2026

#agents #docker #fastmcp #llm #mcp #mcp-server #open-source #python

Overview

PRIMS is a lightweight, open-source Model Context Protocol (MCP) server that safely executes arbitrary Python code in a secure, throw-away sandbox. It runs as a Python-based MCP server, letting LLM agents run code without persistent side effects.

Best for

Best for
Developers who need a minimal, auditable sandbox for LLMs to run Python code without risk.

Use cases

  • Sandboxing AI-generated Python scripts for review
  • Enabling LLMs to run calculations or data transformations safely
  • Testing untrusted code snippets in a contained environment

Notes

PRIMS is a lightweight, open-source Model Context Protocol (MCP) server that safely executes arbitrary Python code in a secure, throw-away sandbox. It runs as a Python-based MCP server, letting LLM agents run code without persistent side effects.

28 stars on GitHub. Last updated 2025-10-15. Licensed MIT.

Use cases

  • Sandboxing AI-generated Python scripts for review
  • Enabling LLMs to run calculations or data transformations safely
  • Testing untrusted code snippets in a contained environment

Pros

  • Simple, single-purpose sandbox with clear security boundaries
  • Open-source and easy to inspect or customize
  • Integrates via standard MCP protocol

Cons

  • Small community and limited real-world testing at 28 stars
  • No built-in resource limits or timeout configurability mentioned
  • Requires manual setup and dependency management

Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.

Pros

  • Simple, single-purpose sandbox with clear security boundaries
  • Open-source and easy to inspect or customize
  • Integrates via standard MCP protocol

Cons

  • Small community and limited real-world testing at 28 stars
  • No built-in resource limits or timeout configurability mentioned
  • Requires manual setup and dependency management