Agent Production Readiness Gate

Scenario

Use after normalizing production cases, running behavioral checks, and auditing operational risks before a release or rollout decision.

Input and output

Input contains case_report, behavior_report, and risk_report objects, plus an optional owner and rollback reference. Output contains decision, blockers, remediation_queue, review_items, and packet_hash.

Execution and recovery

Run the README command. Missing reports or malformed values return a visible error. Any review cases, risk findings, or missing rollback information keeps the decision blocked or human-review-only.

Privacy boundary and acceptance

Use redacted reports and synthetic identifiers. This is a decision aid only; it does not call models, change prompts, grant permissions, deploy agents, or promise production success. The sample creates a non-empty remediation queue and packet hash.