verifying-before-done
Pass
Audited by Gen Agent Trust Hub on Sep 12, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is entirely instructional, focusing on software quality assurance and verification methodologies.
- [SAFE]: It emphasizes credential safety by mandating the redaction of secret values, tokens, connection strings, and authentication headers from any captured output before it is shared.
- [COMMAND_EXECUTION]: The skill encourages running verification commands and capturing their output. This is a core functional requirement for an AI agent performing development tasks and is implemented here as a procedural guideline rather than a hardcoded malicious command.
Audit Metadata