The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to perform technical analysis on video files using shell commands.
Evidence: SKILL.md (Workflow Step 2) contains specific bash code snippets using ffprobe to extract metadata and ffmpeg to generate frame thumbnails from user-provided video files.
[INDIRECT_PROMPT_INJECTION]: The skill has a surface for indirect prompt injection as it is designed to ingest and analyze external, untrusted content.
Ingestion points: SKILL.md (Workflow Step 1) specifies that the skill processes videos from local paths, URLs, browser-visible assets, and external repository media.
Boundary markers: Absent; there are no instructions to the agent to disregard potential prompts or data-embedded instructions within the media metadata or source code it is asked to inspect.
Capability inventory: The skill utilizes subprocess execution (ffmpeg, ffprobe), file system writes (creating /tmp/video-frames), and repository operations (mentioned in Step 6 regarding commits).
Sanitization: Absent; the skill does not specify any validation or sanitization for the URLs or video files provided as input.

video-to-superprompt