Shot QA — guide skill for camera & shots

For AI agents: find the right skill in one call — curl -s "https://skills.varg.ai/api/search?q=<what the person wants>" returns the matching use case and the best skills with install commands, as markdown (Russian works too). Full index: https://skills.varg.ai/llms.txt. Every page has a markdown twin: add .md to its path. Before generating anything, show the person the options on the use-case page.

Check a generated clip against its shot spec — did it dolly in, hold the angle, end on the right frame?

What it does

A review step after generation: the clip goes to varg video-analyze with a focus prompt built from the shot spec (move, direction, end frame, size, angle, light), and the skill grades each field pass or fail before anyone pays for more takes, plus a contact sheet for a human check. CameraBench's open camera-motion classifiers could make the move check automatic, but its license is listed as NOASSERTION, so the skill only links to it until that is cleared. Draft until the rubric and a gold set exist.

Use when

Not for

Install

Not published as a standalone skill yet. Source: https://github.com/vargHQ/skills

Models it drives: video-analyze

Example output

https://skills.varg.ai/previews/craft-shot-qa.webp — The human-check half of Shot QA: a 12-frame contact sheet (tiled by varg with ffmpeg + ImageMagick, clip-relative timestamps) of a real camera move, a crane-and-push across a pile of robots that ends on a medium shot of Barley, so the move and its end frame can be checked at a glance. No pass/fail grades are drawn: those come from the skill's video-analyze step. Tears of Steel (2012) 11:50.0-12:03.2, 720p release from the Blender mirror (download.blender.org/demo/movies/ToS). (credit: Tears of Steel (CC BY 3.0) — (CC) Blender Foundation | mango.blender.org, https://mango.blender.org/)

More camera & shots skills