The Agent Skills Directory

[SAFE]: The skill is a collection of instructional markdown files and local utility scripts (find-polluter.sh, condition-based-waiting-example.ts) intended for use in a development environment.
[COMMAND_EXECUTION]: The skill includes a shell script (find-polluter.sh) used to isolate problematic tests by executing npm test on identified files. This is standard developer tooling. Additionally, SKILL.md provides examples of diagnostic commands (e.g., codesign, security list-keychains) for debugging signing processes, which are used purely for educational and investigative purposes.
[PROMPT_INJECTION]: The provided test files (test-pressure-1.md, etc.) are part of a self-validation framework designed to test the agent's adherence to the debugging process under simulated pressure. These scenarios do not attempt to bypass core AI safety guidelines or extract system prompts.
[CREDENTIALS_UNSAFE]: The skill demonstrates safe practices for handling environment variables by checking for their presence (e.g., ${IDENTITY:+SET}) without printing the actual values.

systematic-debugging