Director-Level Video Lapian

Problem

Video analysis often stops at vague praise such as strong visuals or tight pacing. Creators and engineers need checkable evidence: exact seconds, shot boundaries, sound and subtitle changes, directorial intent, and transferable craft. director-lapian targets films, short videos, animation, ads, music videos, trailers, and scene clips by prioritizing observable facts over invented frame-level claims.

How It Works

The skill starts with a material diagnosis, determining whether the input is video, screenshots, screenplay, storyboard, or link, and stating what precision is possible: frame, second, shot, or paragraph-level inference. It then selects an output level: standard close reading, second-by-second analysis, or ultra-detailed image-text mode. For local video, it preserves a one-second sampling evidence layer while choosing an adaptive report cadence of 0.5, 1, or 2 seconds based on content density. By default, it produces three reports: frame/adaptive lapian report, professional director analysis, and source style bible, plus a vertical showcase video.

Boundaries

When only a screenplay, screenshots, or inaccessible link is provided, the skill states that motion, editing, sound, and true camera facts cannot be verified, and performs a downgraded analysis. It also separates second-level evidence from second-level display: evidence can be retained without printing every frame. It is useful for creative review, style transfer, and prompt preparation, but not for inventing frame-accurate conclusions without source material.