The Agent Skills Directory

[SAFE]: The skill is designed for structural code improvements and explicitly instructs the agent to preserve existing security boundaries, including session checks, role checks, and database interaction logic.
[COMMAND_EXECUTION]: The workflow utilizes php -l for syntax validation. This is a standard development practice for linting PHP files and does not present a security risk in this context.
[PROMPT_INJECTION]: The instructions include protective 'Guardrails' that prevent the agent from making silent behavioral changes or introducing heavy architectures, reducing the risk of unintended behavior or instruction overriding.

refactoring