systematic-debugging

Pass

Audited by Gen Agent Trust Hub on May 14, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill provides a bash script (find-polluter.sh) intended for local use to identify tests that create unwanted state by executing npm test. This is a routine development task.
  • [COMMAND_EXECUTION]: Technical examples in SKILL.md and root-cause-tracing.md describe using standard CLI tools (like env, security, and git) for environment inspection and state verification during debugging sessions.
  • [PROMPT_INJECTION]: The skill contains high-level meta-instructions (e.g., "The Iron Law") and red flags for self-correction. These are process-oriented constraints intended to prevent the agent from taking shortcuts during debugging and do not interfere with system-level safety guidelines.
Audit Metadata
Risk Level
SAFE
Analyzed
May 14, 2026, 01:54 AM
Security Audit — agent-trust-hub — systematic-debugging