Debugging

Pass

Audited by Gen Agent Trust Hub on Jul 3, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill includes scripts/find-polluter.sh, a utility for identifying tests that cause environment pollution. It executes npm test on files matching a provided pattern. This is a standard developer tool pattern and functions as intended for debugging purposes.
  • [PROMPT_INJECTION]: The instructions in references/verification.md and references/systematic-debugging.md use strong prescriptive language (e.g., "Iron Law") to prevent the agent from making unverified claims. This language is focused on ensuring technical accuracy and follows legitimate instructional patterns for agentic workflows rather than attempting to bypass safety filters.
  • [DATA_EXPOSURE]: No hardcoded credentials, API keys, or access to sensitive local configuration files (like .ssh or .aws) were found in the skill's instructions or scripts.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 3, 2026, 01:36 AM
Security Audit — agent-trust-hub — Debugging