effect-uai-model-council

Pass

Audited by Gen Agent Trust Hub on May 4, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill defines a robust architecture for model-to-model evaluation using the reputable 'effect' library for handling asynchronous streams and state management.\n- [SAFE]: All external inputs (model scores) are validated using structured schema decoding ('effect/Schema'), which ensures data integrity and prevents unsafe parsing of model-generated strings.\n- [SAFE]: The skill lacks any dangerous capabilities such as file system access, network exfiltration, or arbitrary command execution.\n- [SAFE]: Error isolation is implemented per-phase, ensuring that failures in individual model turns or scoring do not compromise the overall process or lead to insecure states.
Audit Metadata
Risk Level
SAFE
Analyzed
May 4, 2026, 03:13 PM