Agent skill
shorts-qa
Verify an edited video for complete speech, correct captions, synchronized graphics, audio integrity, image quality and playback.
From timscheuerai/gtm-architecture-template · 51 skill entries · 13 · pushed 2026-09-17
What it does when it runs
Verify an edited video for complete speech, correct captions, synchronized graphics, audio integrity, image quality and playback.
Automated analysis of the skill and the 0 files bundled beside it. A skill’s own description is written to be selected by an agent, so it describes the job and not the dependencies.
- Keys and connectors you must supply
- None found.
- Hosts it reaches
- No third-party host appears in the skill or its bundled files.
- Tool permissions it declares
- No
allowed-toolsin the frontmatter. It only issues instructions, so there is nothing to bound. - Actions present in the files
- None. Instructions only.
Install it
View source on GitHub ↗git clone --depth 1 --filter=blob:none --sparse https://github.com/timscheuerai/gtm-architecture-template.git /tmp/gtm-architecture-template git -C /tmp/gtm-architecture-template sparse-checkout set ".agents/skills/shorts-qa" mkdir -p ~/.claude/skills/shorts-qa cp -R "/tmp/gtm-architecture-template/.agents/skills/shorts-qa/." ~/.claude/skills/shorts-qa/
Picked up without a restart. A project skill of the same name is shadowed by your personal one. For one repository only, swap ~/.claude/skills for .claude/skills. Claude Code docs ↗
The folder is the same in every client that implements the format — 46 of them — so if yours is not above, only the destination changes.
The skill
Source on GitHub ↗Reproduced in full from timscheuerai/gtm-architecture-template/blob/2c312c9398064ad3730c786647293458cb89632c/.agents/skills/shorts-qa/SKILL.md, which is licensed MIT (repository). 224 words, 1 heading.
Verify the final export
Inspect the final encoded artifact against its source and edit plan. No automated seven-check harness is bundled; run available measurements and report what each actually checked.
| Check | Evidence |
|---|---|
| Cuts | Listen to every join; complete words and coherent meaning |
| Captions | Check names/numbers, short words, timing, line breaks and crop |
| Motion | Compare cue/payoff timing while playing audio; inspect overlaps |
| Image | Compare aligned source/output frames outside intentional effects |
| Audio | Listen at matched level; measure peaks/loudness and channel layout |
| Playback | Decode the whole file and open it in the target player when available |
| Pacing | Review the whole story; no duplicate takes or unexplained gaps |
For frame comparisons, align source and output and exclude intended overlays/crops before interpreting quality scores. A threshold from another clip is not universal. Use per-run temporary directories so measurements cannot mix artifacts.
A failed check needs a location, observed defect and fix. Re-check affected portions after correction, then confirm the final file corresponds to the reviewed version. A JSON parse or renderer exit code cannot prove visual/audio quality.
Deliver a short report distinguishing measured, manually inspected and unverified checks. Do not claim a passed gate when the measurement tool or playback inspection was unavailable.
Need help setting it up?
This page tells you what shorts-qa does and what it needs. Cheetah builds the agent setup it runs inside: data, CRM, sequencing and the guardrails.
Book a call →The directory stays free. There is nothing gated behind this.