Video Analysis (Gemini-Native Version)

Use this skill to turn a local video into an auditable analysis package. This skill utilizes a hybrid approach: it first uses local Whisper for robust, high-fidelity transcription to obtain accurate textual context, and then natively leverages Gemini's multimodal video-in capabilities to perform deep content-level analysis, detect semantic scene cuts, and analyze pacing, aligning everything into a readable HTML report.

Default output contract

Create one output folder per source video unless the user specifies otherwise: