agentleFS
nickthelegendnickthelegend/nick-skills/skills/gemini-video/SKILL.md

gemini-video

Skill0 starsChanged 4 days agoMITView on GitHub

Generate short cinematic video clips (hero loops for landing pages, film openers, b-roll for demo videos) with Google's Veo model on gemini.google.com, driven in the user's real Chrome where they are signed in. Covers the prompt, the daily quota, downloading the clip as a blob, encoding it for the web and for films, and keeping Gemini's visible mark. Use it when a landing page or demo film needs a generated motion clip, when someone says "make the video with Gemini", or types /gemini-video.

1 other project uses a version of this.

Adopted by

Most recently taken up by

  • Other versions1 project uses a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 1

    1. austenstone/.copilot★ 0same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repository7 other files it gives its agents.

The file

We split this skill into the rules it gives an agent

6 rules. About 0 are common, 3 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    If Gemini says the limit is reached, report it and move on (use a still from /chatgpt-assets instead).

    line 15
  2. Distinctive:

    Never switch accounts and never try to get around it.

    line 16
  3. Very specific:

    If the source is a blob or MediaSource URL that won't fetch, use Gemini's own download button instead (find it with find).

    line 47
  4. Very specific:

    Use <video autoplay muted loop playsinline poster>, and fall back to the poster under prefers-reduced-motion.

    line 56
  5. Distinctive:

    Keep Gemini's visible sparkle watermark: Don't crop, blur or cover it.

    line 61
  6. Distinctive:

    Never present a generated clip as product footage or as a real event.

    line 63

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

/gemini-video

Veo makes one roughly 8–10 second, 720p clip per request. It's good for a landing-page hero loop or a film's opening beat. It is not for showing the product working: that footage must be a real screen recording of the real product (/judge-skill round 2).

Budget first

  • The user's plan allows very few Veo clips per day. In our runs a second request on the same day was refused. So plan one clip per product per day, and write the prompt carefully before sending.
  • If Gemini says the limit is reached, report it and move on (use a still from /chatgpt-assets instead). Never switch accounts and never try to get around it.

1. Write the shot

One shot, one idea, no text on screen (the model garbles it), no real brands or people. Template:

<camera move>, <subject doing one thing>, <setting>, <light>, <palette as words + hex>, <style>,
8 seconds, seamless loop, no text, no logos, no people's faces

Example: "Slow push-in on a paper lantern slowly lighting up in a dark Tokyo alley at night, warm amber glow (#F5A524) against deep indigo (#14122B), soft film grain, cinematic, 8 seconds, seamless loop, no text, no logos."

Match the product's docs/ART.md (from /chatgpt-assets), so the clip and the stills look like one brand.

2. Generate (Claude in Chrome)

  1. Load the Chrome tools in one ToolSearch call (tabs_context_mcp, tabs_create_mcp, navigate, computer, javascript_tool, find, tabs_close_mcp, browser_batch).
  2. Open https://gemini.google.com/app in a new tab. Pick the video option (Veo) in the composer's tools menu. Take a screenshot, find it, and click it. Insert the prompt with execCommand('insertText', …) and send.
  3. Generation takes a few minutes. Check with a screenshot every minute or so.
  4. Download as a blob from inside the page:

    const v = [...document.querySelectorAll('video')].at(-1);
    const blob = await (await fetch(v.currentSrc || v.src)).blob();
    const a = Object.assign(document.createElement('a'), { href: URL.createObjectURL(blob), download: '<product>-hero.mp4' });
    document.body.appendChild(a); a.click(); a.remove();
    `${blob.type} ${blob.size}`

    If the source is a blob or MediaSource URL that won't fetch, use Gemini's own download button instead (find it with find).

3. Encode

  • Web hero loop (muted autoplay):
    ffmpeg -i in.mp4 -an -vf "scale=1280:-2,fps=30" -c:v libx264 -crf 26 -preset slow -pix_fmt yuv420p -movflags +faststart hero.mp4
    Also make a hero.webm (-c:v libvpx-vp9 -crf 36 -b:v 0) and a poster frame (-frames:v 1 hero.jpg). Use <video autoplay muted loop playsinline poster>, and fall back to the poster under prefers-reduced-motion.
  • Film b-roll: scale to 1920×1080 only in the film's own render. Never upscale the web file.

4. Honesty

  • Keep Gemini's visible sparkle watermark. Don't crop, blur or cover it.
  • Credit it in the README and in the film's end card, for example "Opening clip generated with Gemini (Veo)".
  • Never present a generated clip as product footage or as a real event. A hero loop is mood, not proof.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.