The Agent Skills Directory

[SAFE]: The primary content of the skill is instructional markdown defining a 'Systematic Debugging' methodology. It contains no hidden code, obfuscation, or malicious instructions.
[COMMAND_EXECUTION]: The file find-polluter.sh is a shell script designed to find tests that leak state. It executes npm test on local files matching a user-defined pattern. The script is a standard utility for developer environments and does not present a security risk.
[PROMPT_INJECTION]: The repository contains several 'Pressure Test' files (test-pressure-1.md, etc.). These are benchmark scenarios used to evaluate the AI's adherence to the debugging process. While they simulate adversarial pressure (e.g., 'FIX IT NOW'), they are part of the skill's testing suite and do not represent a threat to the agent's operational safety.

systematic-debugging