# Cisco Launches DefenseClaw Open-Source Framework to Secure AI Agents Across the Enterprise

*2026-03-25 09:02 UTC · Technology · 1 min read*

> Cisco launched DefenseClaw at RSA 2026, an open-source framework securing AI agents with vulnerability scanning, permission controls, and Splunk integration.

Cisco unveiled DefenseClaw, an open-source security framework for AI agents, at RSA Conference 2026 in San Francisco on March 23. The tool bundles vulnerability scanning, permission management, malicious code detection, and Splunk telemetry into a single framework that installs in five minutes and can block rogue MCP servers in two seconds without restarting agents.

The announcement comes as 85% of major enterprises experiment with AI agents but only 5% have moved them into production — a gap Cisco attributes largely to security concerns. "AI agents are not just making existing work faster; they are a new workforce of co-workers that dramatically expand what organizations can accomplish," said Jeetu Patel, Cisco President and Chief Product Officer.

Alongside DefenseClaw, Cisco extended zero trust access to AI agent identities through Duo IAM and Secure Access SSE, and released AI Defense: Explorer Edition, a free self-serve red teaming tool that lets developers test agent resilience before deployment. The framework integrates with NVIDIA OpenShell sandbox project released earlier this month.

## Sources

- [newsroom.cisco.com](https://newsroom.cisco.com/c/r/newsroom/en/us/a/y2026/m03/cisco-reimagines-security-for-the-agentic-workforce.html)
- [siliconangle.com](https://siliconangle.com/2026/03/23/cisco-debuts-new-ai-agent-security-features-open-source-defenseclaw-tool/)

---
TechTools News · https://techtools.cz/tools/news/article/1037-cisco-launches-defenseclaw-open-source-framework-to-secure-ai-agents-across-the
Česká verze: https://techtools.cz/tools/news/clanek/1073-cisco-predstavilo-open-source-framework-defenseclaw-pro-zabezpeceni-ai-agentu-v
