The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/analyze_material.py uses the subprocess module to call external binaries ffprobe and ffmpeg. These calls are used to extract metadata and video frames from user-provided files. The arguments are passed as lists, which is a secure practice that prevents shell injection.
[EXTERNAL_DOWNLOADS]: The documentation in references/image-sourcing.md instructs the agent to use curl for downloading image assets from well-known and legitimate stock photography platforms such as Unsplash, Pexels, and Pixabay. This behavior is consistent with the skill's stated purpose of content creation.
[INDIRECT_PROMPT_INJECTION]: The skill features a data ingestion surface where it processes user-provided text and media files. It mitigates potential issues through structured workflows, specific formatting requirements in references/output-spec.md, and an automated validation script scripts/validate_meta.py that checks the final output against a schema.
[DYNAMIC_EXECUTION]: The workflow relies on external rendering capabilities (Playwright) via a referenced skill card-xiaohongshu to transform HTML templates into final image cards. This is a standard architectural pattern for this type of agent skill.

xhs-note-creator