make-skill
Pass
Audited by Gen Agent Trust Hub on Mar 30, 2026
Risk Level: SAFE
Full Analysis
- Security Best Practices: The skill instructions and associated anti-patterns documentation provide comprehensive security guidance for skill authors. This includes mandates to avoid hardcoding secrets, warnings against unvalidated inputs in shell commands, and the requirement for human oversight on high-stakes actions like PR reviews.
- Fail-Closed Logic: The skill promotes a 'fail-closed' approach for script development (e.g., treating unknown API states as unhealthy), which prevents false positives and improves the integrity of automated workflows.
- Structured Testing Methodology: It defines a multi-model testing process designed to identify edge cases and logic gaps by comparing outputs across different AI model families, which is a proactive measure for ensuring skill safety and correctness.
- Resource Management: The documentation provides sound advice on resource handling, such as disposing of HTTP clients and avoiding temporary file clutter, which contributes to the overall stability and safety of the agent environment.
Audit Metadata