verification-before-completion

Pass

Audited by Gen Agent Trust Hub on Aug 11, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill defines a process for verifying work using industry-standard development tools. No malicious patterns, data exfiltration, or unauthorized access attempts were identified.
  • [COMMAND_EXECUTION]: The skill includes instructions for running common development commands like bun test, npm test, and tsc --noEmit. These commands are limited to their intended purpose of verifying project builds, tests, and linting status within a development environment.
  • [PROMPT_INJECTION]: The skill uses strong imperative language (e.g., "The Iron Law," "Non-negotiable") to establish behavioral constraints. These instructions are designed to prevent the agent from making false success claims and do not target the underlying safety or system instructions of the AI platform.
  • [DATA_EXPOSURE]: No sensitive file paths, credential patterns, or network exfiltration techniques were detected. The skill focuses on standard shell output related to build and test results.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 11, 2026, 09:44 PM
Security Audit — agent-trust-hub — verification-before-completion