analyze-video（分析视频）

Skill: Analyze Video (parent brief)

Adds visual descriptions to a video's audio transcript by extracting JPG frames with ffmpeg and analyzing them.

SKILL.md is the parent's dispatch brief. The sub-agent's working prompt lives in agent_prompt.md — inline its contents when launching the Task agent. Don't pass SKILL.md.

Prerequisites

Each video must already have an audio transcript. Run transcribe-audio first if any are missing.

Parallelism

Launch at most 8 in parallel. ffmpeg frame extraction is a brief CPU burst at the start; the rest of the runtime is LLM API calls. 8 is a comfortable middle ground that won't saturate older machines.

Inputs to gather and pass inline

• video_path — absolute path to the video file
• audio_transcript_path — absolute path to the prepared audio transcript JSON
• visual_transcript_path — absolute path to write the visual transcript JSON

After the agent returns, update library.yaml with visual_transcript: <filename>.json.

Next step

Once all videos have visual transcripts, dispatch summarize-video (Haiku model) to produce summaries.