critique-loop
Pass
Audited by Gen Agent Trust Hub on Jul 13, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill operates as a purely instructional orchestration layer for other established skills (e.g., dex, critique-to-dex). It does not contain executable code, shell scripts, or direct network operations.
- [SAFE]: No patterns for credential exfiltration, prompt injection, or obfuscation were identified across any of the provided files.
- [SAFE]: The skill implements robust procedural guardrails, including 'Thrash Detection' (stopping if regressions exceed fixes) and a default 'Loop Budget' (maximum 3 rounds) to prevent infinite loops or unintended autonomous behavior.
- [SAFE]: The autonomy model is 'opt-in' and requires the agent to confirm progress at round boundaries by default, ensuring human oversight during the workflow.
Audit Metadata