The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes a bash utility find-polluter.sh designed to help developers identify tests that leave behind unwanted state. The script iterates through local test files and executes them using npm test. This is a standard diagnostic operation intended for use within a controlled development environment.
[REMOTE_CODE_EXECUTION]: No remote code execution patterns were detected. The skill provides instrumentation examples for logging environment variables and system state during build processes, which are benign diagnostic commands.
[DATA_EXFILTRATION]: There are no network-capable operations within the provided scripts or instructions. Data remains within the local execution environment for debugging purposes.
[PROMPT_INJECTION]: The skill does not contain any malicious instructions aimed at overriding agent safety protocols or extracting system prompts. It serves purely as an instructional framework to prevent the agent from making hasty or unverified code fixes.

systematic-debugging