Garak-MCP
If you are the rightful owner of Garak-MCP and would like to certify it and/or have it hosted online, please leave a comment on the right or send an email to henry@mcphub.com.
A lightweight MCP (Model Context Protocol) server for Garak LLM Vulnerability Scanner.
The MCP Server for Garak LLM Vulnerability Scanner is designed to facilitate the testing and analysis of language models for vulnerabilities. It provides a structured protocol for interacting with various model types, including ollama, openai, huggingface, and ggml. The server allows users to list available models and probes, run specific attacks, and retrieve reports on vulnerabilities found. This tool is essential for researchers and developers aiming to enhance the security and robustness of language models by identifying potential weaknesses through systematic probing.
Features
- Supports multiple model types: ollama, openai, huggingface, ggml.
- Provides tools to list models and probes for targeted attacks.
- Generates detailed reports on vulnerabilities found during attacks.
- Facilitates integration with various platforms like Claude Desktop and Cursor.
- Allows for extensibility with future support for additional AI models.
Tools
list_model_type
ggml
list_model
List all available models for the specified model type
list_garak_probe
Attack/detection party
get_repor
Get the last run report
run_attac
Run the attack using specified models and detection methods
