The Agent Skills Directory

[SAFE]: The skill is a collection of instructional documents and utility scripts designed to guide an AI agent through a systematic debugging process. It focuses on logical reasoning, data tracing, and validation techniques.
[COMMAND_EXECUTION]: The skill includes a shell script (find-polluter.sh) and various code snippets that suggest executing local development tools (e.g., npm test, git init, security list-keychains, codesign). These commands are standard for technical troubleshooting and diagnostic instrumentation in a software development environment. They do not involve downloading or executing code from untrusted remote sources.
[CREDENTIALS_UNSAFE]: Debugging examples in SKILL.md and root-cause-tracing.md demonstrate how to verify if environment variables or keychain identities are correctly configured. While these touch on sensitive areas (like CI/CD secrets), they are presented as diagnostic steps to identify where configuration is lost in a pipeline, rather than containing hardcoded credentials or exfiltration logic.

systematic-debugging