The Agent Skills Directory

[PROMPT_INJECTION]: The instructions provide guidance on persona and plan structure. No markers for safety bypass or instruction overrides are present.
[DATA_EXFILTRATION]: No network operations or access to sensitive local files such as credentials or configuration are detected. Output is directed to local documentation directories.
[REMOTE_CODE_EXECUTION]: No remote scripts or external code sources are referenced for execution.
[COMMAND_EXECUTION]: References to shell commands like git and pytest are limited to the content of the generated plans as examples for the developer and are consistent with the skill's purpose.
[SAFE]: The skill's functionality is limited to generating implementation documentation and follows standard development practices without suspicious behavior.

Writing Plans