Get the FREE Ultimate OpenClaw Setup Guide →

Mcpwn

mcp security tester

Installation
Run this command in your terminal to add the MCP server to Claude Code.
Run in terminal:
Command
claude mcp add --transport stdio teycir-mcpwn python mcpwn.py

How to use

Mcpwn is an automated security scanner for MCP servers. It analyzes the tooling exposed by MCP implementations to detect command injection (RCE), path traversal, prompt injection, and protocol-related vulnerabilities, and can produce structured reports (JSON/SARIF) suitable for AI analysis and CI/CD pipelines. You can run scans against any MCP server by invoking Mcpwn and pointing it at the target server’s tooling arguments; Mcpwn will perform semantic and side-channel checks, accumulate findings, and categorize them by severity. The tool supports different operation modes, including quick scans, RCE-only mode, safe mode to skip destructive tests, and production security profiles via a paranoid.json profile for stricter testing.

Usage options include: scanning a specific MCP server tool chain, controlling timeouts, enabling parallel testing, and exporting reports in JSON, HTML, or SARIF formats. For example, you can run a basic scan against a server-filesystem tool by executing the Mcpwn CLI with the target payloads, or generate a SARIF report suitable for CI/CD pipelines.

Key capabilities you get with Mcpwn:

  • Semantic detection of RCE patterns, file reads, and timing-based weaknesses in tool arguments.
  • Side-channel detection (timing, size, behavioral anomalies) to reveal non-crash vulnerabilities.
  • Production-security profiling via configurable profiles (e.g., paranoid.json) to apply stricter checks.
  • Structured reporting with severity aggregation and SARIF export for integration with security tooling.
  • Safe mode to skip destructive tests like protocol fuzzing, with options for quick scans and custom timeouts.

How to install

Prerequisites:

  • Python 3.8+ (pure Python / stdlib)
  • No external dependencies required

Installation steps:

  1. Clone the repository: git clone https://github.com/Teycir/Mcpwn.git cd Mcpwn

  2. Ensure the script is executable (optional on some platforms): chmod +x mcpwn.py

  3. Run a test/help to verify installation: python3 mcpwn.py --help

  4. (Optional) Create a virtual environment and install any local development dependencies if needed during customization: python3 -m venv venv source venv/bin/activate

    Install dependencies if you modify the project (not required for stock stdlib-based run)

Notes:

  • The project is designed to run with Python 3.8+ using the standard library; there are no runtime dependencies to install for baseline usage.

Additional notes

Tips and common considerations:\n- Use --profile profiles/paranoid.json to enable production-security-like checks that include side-channel detection.\n- Use --output-json, --output-html, and/or --output-sarif to produce machine-readable reports suitable for CI/CD pipelines.\n- In safe mode, destructive tests such as protocol fuzzing are skipped. This is useful when evaluating a live MCP server.\n- For quick scans, use --quick to limit scan time and stop on the first tool injection finding.\n- If you encounter timeouts, you can raise the timeout with --timeout to accommodate slower targets.\n- If you’re integrating Mcpwn into automation, capture the SARIF output and feed it into your security workflow.\n- The documentation and help output (python mcpwn.py --help) describe all available flags and modes.

Related MCP Servers

Sponsor this space

Reach thousands of developers