The Agent Skills Directory

[SAFE]: The skill establishes a rigorous four-phase debugging methodology centered on identifying root causes before attempting fixes, which reduces the risk of introducing unstable or insecure code.
[COMMAND_EXECUTION]: Includes a shell script (references/find-polluter.sh) that automates the bisection of test files to identify those responsible for unintended file creation or state changes using standard npm commands.
[DATA_EXPOSURE]: Provides educational examples for debugging multi-component systems, including commands to inspect environment variables and system keychains on macOS. These are documented as diagnostic steps for troubleshooting specific build and signing failures.
[SAFE]: Promotes structural security improvements through the 'Defense-in-Depth Validation' reference, which instructs the agent to implement validation checks at multiple layers (entry points, business logic, and environment guards).

systematic-debugging