The Agent Skills Directory

[SAFE]: The skill is a well-structured debugging methodology consisting of instructional Markdown files and helper scripts.- [COMMAND_EXECUTION]: Includes a utility script find-polluter.sh which executes local tests using npm test to identify state-polluting tests. This is a standard development tool functionality.- [DATA_EXPOSURE]: Provides diagnostic examples that inspect environment variables and system keychains (e.g., security find-identity). Notably, the examples include safety-conscious patterns such as checking for the presence of a variable (SET/UNSET) rather than printing its value directly.- [PROMPT_INJECTION]: Uses strong imperative language ('The Iron Law', 'MUST', 'ALWAYS') to enforce a specific logical workflow for debugging. This is an instructional technique for behavior alignment, not a malicious attempt to bypass platform safety guidelines.- [REMOTE_CODE_EXECUTION]: No remote downloads or execution of untrusted external scripts were found. All scripts and commands are intended for local execution within the developer's environment.

systematic-debugging