verification-before-completion

Pass

Audited by Gen Agent Trust Hub on Mar 3, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill defines a strict verification protocol requiring agents to provide evidence (command output) before reporting success. No malicious patterns or vulnerabilities were identified during the analysis.- [PROMPT_INJECTION]: There are no instructions aimed at bypassing safety filters or overriding system prompts. The text focuses entirely on task completion verification logic.- [REMOTE_CODE_EXECUTION]: While the skill discusses running commands, it does not provide or fetch external scripts. It describes a methodology for using existing local tools (tests, linters) to confirm code correctness.- [CREDENTIALS_UNSAFE]: No hardcoded credentials, API keys, or sensitive environmental paths are present in the skill content.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 3, 2026, 08:07 PM