gooseworks-ai/goose-skills / skills/video/video-production-harness
video-production-harness
VerifiedFull video production sequence with script review, actual ingredient choices, controlled generation, editing, evidence-based quality review, polish, captions and delivery. A host binding supplies project storage, authentic approvals, provider access and billing.
references/genres/product-demo.md
View on GitHub# Product demo
A creator explains while real footage proves each claim, or proof points sit over one product shot. Routing rows are named as in the shot-routing guide. Choose speech through [speech and route qualification](../speech-and-route-qualification.md) and size the product through [product scale and world](../product-scale-and-world.md). A text card or dots never replace a live demonstration the reference has. A narrator over several people is [narrator-montage](narrator-montage.md), not this card.
## Shot list skeleton
1. 0-3 s hook: creator full frame, one sharp claim, no footage.
2. 3-15 s use flow, split screen: footage on top, creator below; each line names what shows.
3. 15-20 s payoff hold: the product working or the result; the longest shot.
4. 20-25 s close: the creator says the CTA and the brand.
5. Proof-point variant (10 s): one product shot, a header from 0 s, points landing about every 1.2 s.
## Routing rows
Talking creator, exact script · UI / screen insert · Product photo overlay · Product in hand / motion from a keyframe · Kinetic type / motion card · End card.
## Pacing
- Speech at the reference's measured rate (2.8-3.2 words/s for creator speech), never over 3.2 on a native H3 take; at most 16 words a line; 15-34 s total.
- Each claim's proof on screen within 1.5 s; the payoff holds 2-4× the average shot.
- Footage plays forward at 0.5-2×; no shot under 0.4 s.
## Nearest template
`split-screen-creator-demo` (physical goods: `overlay-proof-points`)
## Top 5 failures and fixes
- A line doesn't match its footage → write each line from its shot; retime on the heard words.
- AI breaks hands, pours or labels → use real footage or photos; move the camera, not the product.
- Words cut at take joins → join on measured speech; a 0.10 s dissolve or a hard cut.
- Text collides or garbles → captions off while points show; all words in HTML, never in a generated picture.
- An invented score or stat → brand-stated facts only; no proof shot means cut the claim.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| receiving-code-reviewobra/superpowersUse when receiving code review feedback, before implementing suggestions, especially if feedback seems unclear or technically questionable - requires technical rigor and verification, not performative agreement or blind implementation | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| requesting-code-reviewobra/superpowersUse when completing tasks, implementing major features, or before merging to verify work meets requirements | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| using-git-worktreesobra/superpowersUse when starting feature work that needs isolation from current workspace or before executing implementation plans - ensures an isolated workspace exists via native tools or git worktree fallback | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| verification-before-completionobra/superpowersUse when about to claim work is complete, fixed, or passing, before committing or creating PRs - requires running verification commands and confirming output before making any success claims; evidence before assertions always | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |