verification-before-completion
Pass
Audited by Gen Agent Trust Hub on Jul 1, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill contains no malicious code, hardcoded credentials, or unauthorized network activity. It defines a set of operational instructions for agents to follow when verifying their work.
- [COMMAND_EXECUTION]: The skill references the execution of verification commands (e.g., tests, builds, linters). These commands are intended to be identified from the specific project context by the agent and do not include hardcoded malicious payloads.
- [PROMPT_INJECTION]: The skill uses authoritative language (e.g., 'IRON LAW', 'NON-NEGOTIABLE') to enforce its verification logic. While strong, these instructions are aimed at improving task accuracy and procedural integrity rather than bypassing the agent's core safety guidelines.
Audit Metadata