CAIN-42 CAIN StudioFAQ
Why not let the AI decide the fix is safe?
Because "safe" was achieved before today by giving the attacker the power. The proof is arithmetic on sets and graphs; the AI never grades its own work.
Which frameworks?
Any that can be described as agents, delegation edges and messages: LangGraph, AutoGen, CrewAI, OpenAI Agents SDK, A2A.
Does it watch my swarm at run time?
No. It audits the declared topology. Run-time enforcement is MCPGate's job.
FIRME AI CORPORATION · CAIN-42 is a Byzantine trust fabric. Numbers marked "our own check" were run by us, not by an independent party. Nothing on these pages is a certification.