The Agent Skills Directory

[PROMPT_INJECTION]: The skill instructions in SKILL.md and beginner-questions.md are focused on educational distillation and user routing. They do not attempt to override system safety filters or engage in role-play jailbreaking.
[DATA_EXFILTRATION]: No hardcoded credentials or sensitive file paths were found. The skill does not perform any network operations; it operates entirely on bundled local reference files.
[REMOTE_CODE_EXECUTION]: There are no patterns of remote script execution (e.g., curl | bash). The Python scripts provided (search_screenshots.py and build_screenshot_assets.py) are for local metadata indexing and image processing.
[COMMAND_EXECUTION]: The skill uses a local Python script (search_screenshots.py) for searching its own database. The input to this script is used for string scoring and does not flow into any dynamic execution sinks.
[EXTERNAL_DOWNLOADS]: References to external sources are limited to the official project repository on GitHub, which is a standard and safe practice for skill distribution.
[SAFE]: The skill includes extensive health warnings and instructions for the agent to refer users to licensed professionals for actual medical needs, maintaining a clear educational boundary.

nihaixia