verification-before-completion
Pass
Audited by Gen Agent Trust Hub on Apr 15, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill contains only instructional content and process guidelines focused on task verification and integrity. No executable code or sensitive operations are present.
- [PROMPT_INJECTION]: The instructions reinforce behavioral constraints rather than attempting to bypass them. Phrases regarding honesty and replacement are used as motivational markers for the agent's performance standards.
- [DATA_EXFILTRATION]: No network operations or sensitive file access patterns were identified. The skill focuses on internal verification of project-specific commands (tests, builds, linters).
- [COMMAND_EXECUTION]: The skill outlines a workflow for running verification commands, but does not provide specific malicious commands. It relies on the agent's existing tools to execute project-relevant validation scripts.
Audit Metadata