The Agent Skills Directory

[SAFE]: The skill's instructions and documentation promote standard software engineering best practices for debugging and system validation. No malicious patterns, obfuscation, or unauthorized access attempts were detected.
[COMMAND_EXECUTION]: The skill provides a bash utility (find-polluter.sh) that executes local test files to isolate tests that create unwanted state pollution.
Evidence: The script executes npm test "$TEST_FILE" using paths found via a user-provided pattern.
Evidence: The script uses standard shell commands (find, sort, ls) for local file system inspection.
[PROMPT_INJECTION]: The skill utilizes strong instructional language (e.g., "The Iron Law") to enforce its debugging methodology.
Evidence: These instructions are internal to the skill's purpose and do not attempt to bypass agent safety filters or override global system instructions.

systematic-debugging