AgentExploitDB protects AI agents and Model Context Protocol (MCP) tools from prompt injections, tool-jacking, and SSRF cloud metadata theft. Unlike passive scanners, it provides a two-pillar defense: • Drop-in zero-dependency runtime shield (MCPGuardrail) stopping attacks in <1ms. • 1,200+ verified OWASP Top 10 exploit vectors for benchmarking. • 1-Click executive HTML security certificate generator for client deliverables. Open source core available on GitHub.