Browse
→ Infrastructure
→ Cisco DefenseClaw
Cisco DefenseClaw
Open-source security governance framework from Cisco AI Defense for agentic AI systems. Scans AI agent skills and MCP servers for vulnerabilities using a combination of the Cisco AI Defense inspect API, YARA rules, and LLM-as-a-judge to detect malicious MCP tools, prompt injections, tool poisoning, and toxic flows. Includes Skills Scanner, MCP Scanner, AI Bill of Materials generator, and CodeGuard for static analysis. Ensures every agent skill is scanned, every MCP server verified, and every AI asset cataloged automatically. Released at RSA 2026. Apache 2.0 licensed. Python-based.
MCP unverified
Integration
| Transport | stdio |
| Auth | api-key |
| Endpoint | cisco-defenseclaw-mcp |
Use Cases
| 01 | Scan MCP servers and agent skills for security vulnerabilities including prompt injection, tool poisoning, and toxic flows before deploying them in production |
| 02 | Generate AI Bills of Materials cataloging every agent, skill, and MCP server in your organization for compliance and audit purposes |
| 03 | Integrate automated security governance into CI/CD pipelines to verify that every MCP server and agent skill meets security policies before release |
Tags
security scanning governance vulnerability-detection cisco enterprise open-source ai-bom rsa-2026
Machine-readable: /api/servers.json
· JSON-LD schema embedded in <head>