decision-critic
Pass
Audited by Gen Agent Trust Hub on Apr 26, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: Technical review of the skill's Python script and markdown instructions found no evidence of malicious behavior. The script performs string formatting and provides step-by-step guidance to the agent without external dependencies or network access.
- [PROMPT_INJECTION]: The skill ingests user-supplied decision text. 1. Ingestion points: decision-critic.py arguments --decision and --context. 2. Boundary markers: The script uses 'DECISION UNDER REVIEW:' and 'CONTEXT:' headers to delimit content. 3. Capability inventory: The skill is restricted to text analysis and lacks network access, file system write permissions, or shell execution capabilities. 4. Sanitization: No explicit sanitization of input text is performed. The finding is classified as safe due to the lack of exploitable capabilities.
Audit Metadata