The Agent Skills Directory

[DATA_EXFILTRATION]: The skill includes a specific production guardrail that explicitly forbids documenting secrets, private credentials, internal tokens, or sensitive customer data, effectively mitigating accidental credential exposure.
[COMMAND_EXECUTION]: The skill uses standard development commands like git diff and reads project configuration files (e.g., package.json, Makefile, CI config) to verify changes, which is appropriate for its documented purpose.
[PROMPT_INJECTION]: The 'Freshness Check' section defines documentation severity levels (Critical, High, Low) to guide the agent in prioritizing documentation fixes. These terms are used in the context of documentation accuracy and do not attempt to override the underlying model's safety or behavioral constraints.

documentation