Cisco DevNet Adds Live AI Security Labs for MCP, A2A Protocol, and AI Defense
Cisco's DevNet Learning Labs platform has rolled out a new set of hands-on security labs specifically designed for developers working with emerging agentic AI standards. The lineup covers four domains that are rapidly becoming critical as AI deployments mature: MCP Security, A2A Protocol Security, AI Defense, and Skill Security. Each lab is built around the kind of practical, scenario-based learning that is hard to replicate from documentation alone.
One of the more thoughtful design decisions in the new labs is the elimination of setup friction. No LLM provider account is required, and no model configuration needs to happen before a developer can get started — managed LLM access is built directly into the lab environment. That removes a meaningful barrier that typically slows security education, particularly for teams that want to evaluate agentic security postures before they have committed to a specific provider stack.
As MCP and A2A move from experimental specifications toward the communication standards that real multi-agent systems will depend on, the security surface they introduce deserves far more attention than it currently receives. Cisco's decision to build dedicated labs around these protocols — rather than treating them as footnotes in broader AI security curricula — reflects a realistic read of where enterprise AI infrastructure is heading. For developers building or securing agentic systems, these labs are a practical place to start hardening before the first production deployment.