Official MCP server that lets AI tools read repositories, manage issues and pull requests, and query GitHub Actions and security data.
Autonomous API and web-app penetration testing tool that runs a swarm of AI agents to chain exploits and prove vulnerabilities.
A registry and proxy that gives agents one base URL and token to call a catalog of third-party tool endpoints, billed per call.
Open source proxy that routes and secures agent-to-LLM, agent-to-tool, and agent-to-agent traffic using MCP and A2A protocols.
General-purpose toolkit implementing TLS, DTLS, and QUIC protocols on top of a standalone cryptographic library and CLI.