red-team-blue-team-agent-fabric: The Agent Security Harness is a security testing framework for AI agents, offering 470 executable tests across 32 modules covering MCP, A2A, L402, x402 wire protocols, and decision-layer attack scenarios. It provides active protocol exploitation testing beyond static analysis, with research backing and a focus on both protocol integrity and governance.; thunderbit-mcp-server: Thunderbit MCP Server is an open-source toolkit for the Thunderbit Open API that ships three packages: a CLI for scripted extraction, an MCP server exposing seven scraping and distillation tools, and a Claude Code plugin. It converts any web page to clean LLM-ready Markdown, extracts structured data via JSON Schema, and supports batch processing — all backed by a free API key.
Security testing of MCP servers
Feeding clean web content into LLM pipelines for RAG or summarization