verification-before-completion

Pass

Audited by Gen Agent Trust Hub on Mar 1, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is entirely composed of markdown instructions and does not include any scripts, executables, or network-active components.
  • [PROMPT_INJECTION]: While the skill uses high-pressure language (e.g., "The Iron Law", "If you lie, you'll be replaced"), this is directed at enforcing a technical verification workflow for software development rather than attempting to bypass safety filters or manipulate the underlying LLM's core instructions.
  • [DATA_EXPOSURE]: The instructions focus on reviewing local command outputs and version control diffs (VCS), which is standard behavior for a development assistant and does not indicate unauthorized data access or exfiltration.
  • [COMMAND_EXECUTION]: The skill references running standard development commands (tests, builds, linters), but it does not provide specific or obfuscated command strings to execute. It serves as a behavioral guide for the agent's existing capabilities.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 1, 2026, 08:41 PM