The Agent Skills Directory

[SAFE]: The skill serves as an internal process guide for the agent to conduct post-response verification. It focuses on 'self-improvement' by instructing the agent to run adversarial evaluations on its own design or implementation work after the primary response is delivered. It references other internal skills (e.g., kf-g-agent-review-adversarial-evaluation) for specific evaluation formats but does not contain executable code, external network calls, or sensitive data access. The instructions emphasize waiting for an initial response to be sent before starting the review process, ensuring a clear boundary between delivery and verification.

kf-g-agent-review-adversarial-after-primary-response