ai Agent Production Hardening Kit

Getting an AI agent to demo is easy. Getting it to survive real users isn't. This skill walks every production risk: prompt injection attack surfaces, jailbreak patterns, tool-call blast radius, hallucination failure modes, retry and circuit-breaker logic, observability hooks, and graceful degradation paths. Returns a threat model specific to your agent architecture, a guardrail specification ready to implement, and a production readiness scorecard you can actually ship against. Whether you're hardening a customer-facing agent or an autonomous pipeline, this is the pre-launch checklist you can't afford to skip.