Rlm Mcp — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited Rlm Mcp (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
MCP (Model Context Protocol) server wrapper for RLM (Recursive Language Models).
Note: This is an MCP interface for the RLM library. The core RLM implementation is by Alex Zhang, Tim Kraska, and Omar Khattab at MIT CSAIL. See Acknowledgments for full credits.
RLM enables verified code execution with LLM reasoning - it writes and executes Python code iteratively until producing a verified answer.
git clone https://github.com/alexzhang13/rlm.git $HOME/rlm
cd $HOME/rlm
pip install -e . export OPENROUTER_API_KEY="your-key-here"# Create MCP server directory
mkdir -p $HOME/.claude/mcp-servers/rlm
# Download files
curl -o $HOME/.claude/mcp-servers/rlm/src/server.py \
https://raw.githubusercontent.com/eesb99/rlm-mcp/main/src/server.py
curl -o $HOME/.claude/mcp-servers/rlm/run_server.sh \
https://raw.githubusercontent.com/eesb99/rlm-mcp/main/run_server.sh
curl -o $HOME/.claude/mcp-servers/rlm/setup.sh \
https://raw.githubusercontent.com/eesb99/rlm-mcp/main/setup.sh
curl -o $HOME/.claude/mcp-servers/rlm/requirements.txt \
https://raw.githubusercontent.com/eesb99/rlm-mcp/main/requirements.txt
# Setup
chmod +x $HOME/.claude/mcp-servers/rlm/*.sh
$HOME/.claude/mcp-servers/rlm/setup.shAdd to $HOME/.mcp.json:
{
"mcpServers": {
"rlm": {
"command": "bash",
"args": ["/YOUR/HOME/PATH/.claude/mcp-servers/rlm/run_server.sh"]
}
}
}Replace /YOUR/HOME/PATH with your actual home directory (run echo $HOME to find it).
| Variable | Default | Description |
|---|---|---|
OPENROUTER_API_KEY | (required) | OpenRouter API key |
RLM_MODEL | openrouter/x-ai/grok-code-fast-1 | Root execution model |
RLM_SUBTASK_MODEL | openrouter/openai/gpt-4o-mini | Subtask model |
RLM_MAX_DEPTH | 2 | Max recursion depth |
RLM_MAX_ITERATIONS | 20 | Max iterations per task |
RLM_LOG_DIR | ~/.rlm/logs | Directory for execution logs |
RLM_LIB_PATH | $HOME/rlm | Path to RLM library (if not pip installed) |
# Install mcporter
npm install -g mcporter
# Check server is available
mcporter list | grep rlm
# Execute a calculation
mcporter call 'rlm.rlm_execute(task: "calculate the first 20 prime numbers")'
# Analyze data
mcporter call 'rlm.rlm_analyze(data: "[1,2,3,4,5]", question: "what is the mean?")'
# Check status
mcporter call 'rlm.rlm_status()'RLM executes arbitrary Python code by design. Only use with trusted inputs. The code runs in a local Python environment without additional sandboxing.
This MCP server is a wrapper for the Recursive Language Models (RLM) library developed by:
The RLM concept and implementation are their original work. This repository only provides an MCP interface to make RLM accessible via the Model Context Protocol.
Citation:
@article{zhang2025rlm,
title={Recursive Language Models},
author={Zhang, Alex L. and Kraska, Tim and Khattab, Omar},
journal={arXiv preprint arXiv:2512.24601},
year={2025}
}MIT License - see LICENSE
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.