cavekit-validation-first

Pass

Audited by Gen Agent Trust Hub on Jun 13, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill provides instructional guidance on software development validation procedures. It outlines a structured methodology for ensuring that AI-generated code meets technical requirements through automated and manual gates.
  • [COMMAND_EXECUTION]: The skill references various command templates (e.g., {BUILD_COMMAND}, {TEST_COMMAND}, {E2E_TEST_COMMAND}) intended to be populated with project-specific commands. These are standard development operations and do not include hardcoded malicious payloads.
  • [DATA_EXFILTRATION]: Network activity described is limited to checking the health of a local application instance (http://localhost:{PORT}/health) during smoke testing. This is a standard practice and does not involve sending data to external or untrusted domains.
  • [PROMPT_INJECTION]: The instructions focus on improving agent reliability through validation criteria. There are no attempts to override safety filters, extract system prompts, or bypass platform constraints.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 13, 2026, 01:03 AM
Security Audit — agent-trust-hub — cavekit-validation-first