What problem does it solve? Reviewing video content normally requires watching or listening to the entire recording, which is slow and disruptive. This Skill answers questions about videos by reading transcripts first and inspecting footage only when visuals, sound, or timing actually matter, all without audible playback. ## Core Features & Use Cases - Transcript-First Routing: Summarize arguments, extract quotes, and repurpose content from timestamped transcripts, escalating to video inspection only for ambiguous or visual questions. - Silent Footage Verification: Check demonstrations, delivery, editing, and clip boundaries using a local tool or authorized Gemini API access without playing sound on the user's machine. - Clip Boundary Validation: Verify proposed clip openings and endings against timestamped audio/video evidence, labeling model-estimated timestamps as approximate. - Use Case: Ask whether a screen demo in a YouTube tutorial supports the speaker's claim; the Skill reads the transcript, inspects only the relevant video ranges via Gemini, and returns a grounded answer with timestamp links. ## Quick Start Use the video-analysis skill to summarize this YouTube video's argument from its transcript and verify any on-screen claims silently.
