mcp-server-code-execution-mode

🥈Silver

An MCP server that executes Python code in isolated rootless containers with optional MCP server proxying. Implementation of Anthropic's and Cloudflare's ideas for reducing MCP tool definitions context bloat.

321240Updated 1w ago

Intermediate30min to implementautomation

Saves ~180 min per use

Quick InstallView Source

git clone https://github.com/elusznik/mcp-server-code-execution-mode.git

Works with:

Claude

Overview

About This Skill

How to Use

["Install the MCP server: `mcp install mcp-server-code-execution-mode` (requires Docker or Podman for containerization)","Prepare your Python code in a file (e.g., `script.py`) or paste it directly into the prompt. Include all dependencies in the prompt or as a requirements.txt file.","Execute with context: Add `[INPUT_DATA]` for stdin input or `[OUTPUT_FORMAT]` to specify JSON/YAML output. Example: `Use the server to run [script.py] with [INPUT_DATA='{\"filter\": \"West\"}'] and return [OUTPUT_FORMAT='json']`","For complex workflows, chain executions by referencing previous outputs: `Take the output from [PREVIOUS_EXECUTION] and run [script.py] with it as [INPUT_DATA]`","Monitor resource usage: The server provides real-time memory/CPU metrics. For long-running tasks, add `[TIMEOUT=300]` to limit execution time."]

Setup & Installation

Quick Install

No install command available. Check the GitHub repository for manual installation instructions.

Alternative Install (Git Clone)

git clone https://github.com/elusznik/mcp-server-code-execution-mode

Requirements

Claude Code or compatible AI agent
Works with: Claude

Quick Start Guide

Install the Skill

Copy the install command above and run it in your terminal.

Open Your AI Agent

Launch Claude Code, Cursor, or your preferred AI coding agent.

Try It Out

Use the prompt template or examples below to test the skill.

Customize

Adapt the skill to your specific use case and workflow.

Usage Examples

Prompt Template

Use the mcp-server-code-execution-mode to execute the following Python code in an isolated container: [PYTHON_CODE]. If the code requires external dependencies, specify them as [DEPENDENCIES]. For complex workflows, chain multiple executions with [INPUT_DATA] and [OUTPUT_FORMAT].

Example Output

```python
# Example: Data Processing Pipeline
import pandas as pd
import numpy as np

# Input data (simulated)
sales_data = pd.DataFrame({
    'date': pd.date_range(start='2023-01-01', periods=100),
    'product_id': np.random.choice(['A100', 'B200', 'C300'], 100),
    'revenue': np.random.uniform(1000, 5000, 100).round(2),
    'region': np.random.choice(['North', 'South', 'East', 'West'], 100)
})

# Data processing
processed = (
    sales_data
    .groupby(['product_id', 'region'])
    .agg({'revenue': ['sum', 'mean', 'count']})
    .reset_index()
)

# Output
processed.columns = ['product_id', 'region', 'total_revenue', 'avg_revenue', 'transaction_count']
processed['total_revenue'] = processed['total_revenue'].round(2)
processed['avg_revenue'] = processed['avg_revenue'].round(2)

print("Top 5 Revenue-Generating Product-Region Pairs:")
print(processed.nlargest(5, 'total_revenue').to_string(index=False))
```

**Execution Log:**
- Container: Ubuntu 22.04 (rootless, 2GB RAM)
- Dependencies: pandas==2.0.3, numpy==1.24.3 (installed via pip in 8s)
- Runtime: 1.2s
- Memory Usage: 142MB peak

**Output:**
```
Top 5 Revenue-Generating Product-Region Pairs:
 product_id region  total_revenue  avg_revenue  transaction_count
      A100   West        24567.89     2456.79               10
      B200   North       22345.67     2234.57               10
      C300   South       21890.12     2189.01               10
      A100   East        20123.45     2012.35               10
      B200   West        19876.54     1987.65               10
```

**Security Notes:**
- Filesystem: Ephemeral (no persistence after execution)
- Network: Disabled by default (enable with --allow-network if needed)
- User: Non-root (UID 1000)