Communitygithub.com

SingleGrain-Team/video-skills

Build a 15-20 second product launch film that shows the product's real screen on a laptop built from plain CSS (no 3D model, no generation model) in a dark studio, scrolls the page, zooms into the interface for one key action such as typing a search, and resolves to the logo lockup. Use when someone has a full-page screenshot or screen recording of their product, one action worth showing, and a logo, and wants a short silent-friendly launch teaser rendered as a 1920x1080 MP4 with an editable HyperFrames project.

Was ist video-skills?

video-skills is a Claude Code agent skill that build a 15-20 second product launch film that shows the product's real screen on a laptop built from plain CSS (no 3D model, no generation model) in a dark studio, scrolls the page, zooms into the interface for one key action such as typing a search, and resolves to the logo lockup. Use when someone has a full-page screenshot or screen recording of their product, one action worth showing, and a logo, and wants a short silent-friendly launch teaser rendered as a 1920x1080 MP4 with an editable HyperFrames project.

Funktioniert mit~Claude Code~Codex CLI~Cursor
npx skills add https://github.com/SingleGrain-Team/video-skills/tree/HEAD/skills/device-launch-film

In Ihrer bevorzugten KI fragen

Öffnet einen neuen Chat, in dem dieser Agent-Skill bereits geladen ist.

Dokumentation

device-launch-film

A single-shot launch film: device appears, the real page scrolls, the camera pushes into one action, the device falls away and the logo lockup lands. About 18 seconds, no narration required.

Inputs

Get these before building. Ask for any that are missing.

  1. Page capture: a full-page PNG of the real product (preferred, 1920 wide, up to roughly 6000 tall) or a screen recording. For a recording, extract a still of the starting state and use that for the scroll; use the clip itself only for the zoom beat.
  2. The one action: a single interaction that reads clearly in 3 seconds. Typing into a search field is the easiest; a button press or toggle also works. Note its location in page pixels.
  3. Logo: transparent PNG or SVG, plus product name and a one-line tagline. Use only assets the user owns.
  4. Optional: brand colors (accent, dark background), URL for the closing line.

Never substitute a third-party UI or logo for the product's own.

Workflow

  1. Load /hyperframes and /hyperframes-core; follow their composition contract.
  2. Scaffold a project directory. Put the capture in assets/page.png and the logo in assets/logo.png.
  3. Copy template/index.html from this skill as index.html. It is one standalone composition driven by a CFG block and a handful of CSS sizes.
  4. Fit the page to the screen. The screen is 1120x700 (16:10) inside the lid. Scale = 1120 / capture width. Set CFG.scale and CFG.maxScroll (= capture height x scale minus 700). If the capture is shorter than the screen, drop the scroll beat; if it is very tall, keep the scroll to 3-4 seconds and let it be fast.
  5. Locate the action in the capture (page px) and write down the field's box and the text origin. In the template this means editing the #patch (covers the placeholder), #ring (focus outline), #typed (text position and font size) rules, and CFG.focus (the stage point the camera centers on: screen left 400 + x * scale, screen top 110 + y * scale, with the scroll back at 0).
  6. For a result state, either crop the real result from a second capture, or reuse a real element from the same capture the way the template does (#hit is a cropped card slid under the field while #cover hides the rest). Never draw invented data.
  7. Replace the lockup text and colors in #end. Adjust the timeline only through the numbers already in it.
  8. Run npx --yes hyperframes check and fix every warning you caused (see Validation), then npx --yes hyperframes snapshot --at 0.8,3.5,6,9,10.5,12,13,16 and look at the contact sheet.
  9. Render: npx --yes hyperframes render --workers 2 --fps 30 -o final.mp4.

Beat sheet (18 s)

TimeBeat
0-1.6Laptop rises out of the dark; screen brightens from black
1.7-5.2Page scrolls top to bottom with an ease-in-out, camera drifts in 3 percent
5.6-7.1Scroll returns to the top
7.0-8.3Camera zooms (about 3x) onto the action target; focus ring appears
8.7-10.5The action plays (typing at about 5 characters per second)
10.8-12.4The result appears and rises into place
12.7-13.8Camera pulls back, device shrinks and fades
13.9-18Lockup: mark pops in, name, tagline, URL; hold at least 2 seconds

Rules

  • Device is CSS only: lid with bezel, screen, base with a notch, a soft floor shadow. No images of real hardware, no 3D, no trademarked device silhouettes beyond a generic laptop slab.
  • Two nested camera wrappers: the outer one drifts and fades, the inner one zooms. Never tween the same property on the same element from two tweens that overlap in time.
  • Everything on the screen lives inside the scaled page element so the zoom stays sharp and aligned. Overlays (typed text, ring, patch, result) use page pixels.
  • Typing is driven by a proxy number and onUpdate, which stays seek-safe. No timers, no random.
  • Name no web fonts without a local @font-face; the template uses the system stack. If the brand needs a font, ship the file in assets/ and declare it.
  • Do not show real people's photos large. If the capture has avatars, keep the zoom away from them or cover them with the page background.
  • Music is optional. Add it only if the license permits commercial use and record the source; silent is fine.
  • Do not log in, purchase, or capture private data to make the capture.

Validation

  • check reports 0 errors. The layout notes container_overflow on the scroller and escaped_container on the zoomed camera are expected here (the scroller is meant to overflow its clip; the camera is scaled on purpose). Any other warning, especially overlapping_gsap_tweens, must be fixed.
  • Snapshot the key beats and open them: device centered and not cropped at 0.8 s and 3.5 s, the zoomed field readable with its full typed text, the result clear of the field, the lockup centered with nothing left of the device on screen.
  • After rendering, confirm with ffprobe that the file is H.264, 1920x1080, 15-20 s, then extract four to six frames and look at them. If anything is broken, fix and re-render.

Delivery

Hand over final.mp4, the project folder (editable), and a short note listing the capture and logo used, any audio source, and the CLI version. Mention that the project can be edited in HyperFrames Studio.

Individual skills in this repo

This repo contains 7 individual skills — each has its own dedicated page.

SingleGrain-Team/video-skills

Make a 30-45 second launch film for a product by staging it inside one persistent, neutral chat window. A user types a request, the assistant answers and calls the product as a connected tool, a progress card runs, a follow-up triggers a second tool action, and the result lands as a count-up headline number before closing on the logo, tagline and call to action. Use when you need a launch, feature or integration announcement video for something that plugs into AI assistants (a skill, MCP server, plugin, connector, CLI). Delivers a rendered 1920x1080 MP4 plus an editable single-file HyperFrames project.

SingleGrain-Team/video-skills

Build a motion graphic where a screen recording plays in a large frame beside a checklist, and each checklist item ticks off at the exact moment that step happens in the recording, ending with every item checked and a short hold. Use when someone has a screen recording (product walkthrough, tutorial, onboarding, install demo) plus a list of steps with timestamps and wants a "watch the steps get done" film. Input is a recording, step labels with timestamps, a title and an optional palette; output is a 1920x1080 H.264 MP4 and the editable HyperFrames project.

SingleGrain-Team/video-skills

Builds a 6-8 second logo sting in which the brand mark rides inside a refracting liquid bead (distorted past legibility), the bead swells, holds still for a beat, then ruptures into droplets and leaves the logo lockup (mark plus wordmark) crisp and upright on screen. Delivered as a 1920x1080 H.264 MP4 plus the editable HyperFrames project, with no 3D model, stock footage or generation model. Use it for brand intros and outros, product-launch openers, channel stings and any "reveal our logo" moment, from an SVG or PNG logo (including raster-only logos), a wordmark string and a palette.

SingleGrain-Team/video-skills

Make a 20-30 second launch film for a product that is not tied to one interface, such as a CLI, package, plugin or agent skill. The identical prompt or command is typed into four visually different surfaces in turn (a wide terminal, a tall chat card, a floating palette, an editor cell), the film then hard-cuts into a fast reel of real outputs on a beat or fixed tempo grid, and it closes on the command in a large pill. Delivered as a rendered 1920x1080 MP4 plus the editable HyperFrames project. Music is optional - with a track the cuts land on detected beats, without one they land on a 120 bpm grid.

SingleGrain-Team/video-skills

Make a roughly 30-second film that explains a code change (a pull request, release or uncommitted diff) to people who will never read the diff. It opens by showing the real diff scrolling past too fast to follow, cuts to a plain-language summary, fires four capability beats each pinned to a real code snippet, and closes on how to get it (install command and URL). Delivered as a 1920x1080 H.264 MP4 plus the editable HyperFrames project. Use when asked to turn a PR, changelog or release into a launch or explainer video.

SingleGrain-Team/video-skills

Render a 5-6 second motion graphic in which an oversized condensed headline (2-4 words) punches in behind a presenter, using a background-removed cutout of the presenter layered on top of the lettering. A hard cut then removes the presenter while the same headline holds the exact same size and position, and a short supporting line appears for a reading hold. Use it for talking-head intros, chapter openers, launch announcements and social hooks when you have one presenter clip, a headline, a supporting line and a palette. Built on HyperFrames; delivers a 1920x1080 H.264 MP4 plus the editable project.

SingleGrain-Team/video-skills

Builds a 10-14 second product sting out of a short run of UI moments (an entry screen, a single counting metric, a navigation list, one click that opens a menu, then an endcard) cut together so every cut carries the previous motion forward and the piece reads as one continuous camera move. Delivered as a 1920x1080 H.264 MP4 plus an editable single-file HyperFrames project. Use it for app launches, feature teasers, social bumpers and site intros when you have a product's logo, a couple of screenshots, one headline number and its real nav labels, and want polished UI motion without narration.

Verwandte Skills