youtube-thumbnails
Skill194 starsChanged 8 months agoMITView on GitHub
This skill should be used when the user asks to "create YouTube thumbnails", "generate thumbnails", "make video thumbnails", or needs AI-generated thumbnail images via Google Gemini for YouTube videos. Covers prompt engineering, image generation, compression, and upload to YouTube Studio.
5 other projects use a version of this, 1 of them edited.
Adopted by
Most recently taken up by
- nikhilbhansali/youtube-data-skillstheir own version, first seen Nov 8, 2026
- JorgeMenaDev/skillstheir own version, first seen Oct 20, 2026
- Pesty-Marketing/agent-skillstheir own version, first seen Oct 18, 2026
- matthiasloeffel/claude-skillstheir own version, first seen Oct 15, 2026
- biyearly-mesothelioma790/skillsedited, first seen Mar 28, 2026
Other versions5 projects use a version of this.
Word for word · 0
None found.
Edited · 1
- biyearly-mesothelioma790/skills★ 0100% same
Wrote their own version · 4
- nikhilbhansali/youtube-data-skills★ 4same name
- JorgeMenaDev/skills★ 1same name
- matthiasloeffel/claude-skills★ 0same name
- Pesty-Marketing/agent-skills★ 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 repository102 other files it gives its agents.
- AGENTS.mdAGENTS.md
- accelevents-apiSkill
- accelevents-speaker-syncSkill
- agents-sdkSkill
- ai-devblogSkill
- ai-engineeringSkill
- ai-readmeSkill
- align-meSkill
- annotated-video-productionSkill
- antislop-codebaseSkill
- app-ux-paradigmsSkill
- autoreviewSkill
- babysit-runsSkill
- blog-system-designSkill
- cassandra-creatorSkill
- ceo-creatorSkill
- claude-session-introspectSkill
- cli-uxSkill
- cloudflare-do-turn-based-multiplayerSkill
- cloudflare-email-serviceSkill
- cloudflare-one-migrationsSkill
- cloudflare-oneSkill
- cloudflare-production-builderSkill
- cloudflareSkill
- conference-developer-endpointsSkill
- conference-transcribeSkill
- coordinateSkill
- data-chatbotsSkill
- data-visualization-qualitySkill
- deep-trajectory-analysisSkill
- design-apps-with-imagegenSkill
- design-preferencesSkill
- diy-netlifySkill
- download-videoSkill
- download-x-videoSkill
- durable-objectsSkill
- europe-developer-apiSkill
- face-matchingSkill
- forgeSkill
- future-onlySkill
- generalizeSkill
- gsuite-setupSkill
- keep-it-simpleSkill
- latent-space-thumbnail-directorSkill
- live-ai-pipelinesSkill
- location-inputSkill
- long-running-operation-uxSkill
- make-evalsSkill
- media-heavy-workflowsSkill
- media-to-wiki-pipelineSkill
- media-transformSkill
- mobile-nativeSkill
- mobile-webapp-uxSkill
- multimodal-extractionSkill
- new-mac-setupSkill
- next-stepsSkill
- observability-hardeningSkill
- person-profile-writingSkill
- podcast-publishing-assistantSkill
- productionize-app-with-servicesSkill
- The first 60 shown.
The file
We split this skill into the rules it gives an agent
7 rules. 0 are common, 5 are distinctive, and 2 are very specific to this project.
0
Common
Most of the 4 edited versions we read keep these.
5
Distinctive
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 35
Use this prompt template, customized per video:
- Distinctive: line 62
Always ask Gemini to search the web for relevant logos and reference images
- Distinctive: line 66
Keep the background dark so text pops
- Distinctive: line 90
Verify the thumbnail appears in the preview
- Distinctive: line 94
After clicking Upload file, use osascript to navigate the native file dialog:
- Very specific: line 113
After convert -resize 1280x720 -quality 85: ~200-250KB (well under 2MB limit)
- Very specific: line 145
Upload button not visible: Use the find tool to locate the hidden "Upload file" button by ref.
“Common” is measured: kept by at least 6 in 10 of the 4 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
YouTube Thumbnail Generation via Gemini
Overview
Generate custom YouTube thumbnails using Google Gemini's image generation (Pro mode). Each thumbnail is a 1280x720 image with bold text overlays, thematic visuals, and branding.
Step-by-Step Workflow
1. Gather Video Info
For each video that needs a thumbnail, collect:
- Title (already set on YouTube)
- Content type: "AI in Action" weekly jam or "Paper Club" reading
- Topic summary: What was presented/demoed (from frame extraction or title)
- Presenter name (optional, for subtitle text)
- Date in
D Mon YYYYformat
2. Navigate to Gemini
- Go to
gemini.google.com - Click Tools button (bottom toolbar)
- Enable Create images
- Click the model dropdown (shows "Fast" by default) and select Pro
- Confirm "Image" chip and "Pro" label appear in the chat input area
3. Craft the Thumbnail Prompt
Use this prompt template, customized per video:
Create a YouTube thumbnail (1280x720) for a video titled "[FULL TITLE]".
[1-2 sentences describing what the video is about — the demo, paper, or topic].
The visual should be [THEME DESCRIPTION]: [specific visual elements relevant to the topic].
Bold text overlay: "[MAIN TITLE]" prominently, with smaller text "[SUBTITLE]" below it.
Include a bottom bar with "[SERIES NAME] [D MON YYYY]" in bold.
Please include the Latent Space podcast logo in a corner — search the web for "Latent Space podcast logo" to find it.
Use [COLOR PALETTE] with a dark [THEME] background.Prompt Customization by Content Type
AI in Action (demos/presentations):
- Bottom bar text:
AI IN ACTION [D MON YYYY] - Visual theme: Match the demo topic (space, gaming, dev tools, etc.)
- Colors: Vibrant, exciting — neon blues, greens, purples, electric tones
Paper Club (paper readings):
- Bottom bar text:
PAPER CLUB [D MON YYYY] - Visual theme: Academic but engaging — neural networks, brain visualizations, formulas
- Colors: Deep blue, orange, white — scholarly but modern
- Main text: Paper acronym or short name (e.g., "SDPO")
- Subtitle: Full paper name
Key Prompt Tips
- Always ask Gemini to search the web for relevant logos and reference images
- Specify 1280x720 dimensions explicitly
- Request bold, large text for the main title — readable at thumbnail size
- Include a colored bottom bar with the series name and date
- Keep the background dark so text pops
4. Generate and Save
- Submit the prompt and wait ~30-45 seconds for Pro mode generation
- Click the generated image to expand it
- Click Save to download to ~/Downloads
- Generated files will be named
Gemini_Generated_Image_[hash].jpeg
5. Compress to Under 2MB
YouTube requires thumbnails under 2MB. Gemini Pro images are typically 3-4MB.
convert "Gemini_Generated_Image_[hash].jpeg" -resize 1280x720 -quality 85 thumb_[label].jpgNaming convention: thumb_[abbreviated_date].jpg (e.g., thumb_jan31.jpg, thumb_feb7.jpg)
6. Upload Thumbnails to YouTube Studio
For each video:
- Go to YouTube Studio → Content → click on the video
- Click the pencil/edit icon (Details tab) in the left sidebar
- Scroll to the Thumbnail section
- Click the Upload file button (may be hidden — use
findtool to locate ref) - Select the correct
thumb_[date].jpgfile - Verify the thumbnail appears in the preview
- Click Save at the top right
File picker automation (macOS): After clicking Upload file, use osascript to navigate the native file dialog:
tell application "System Events"
delay 1
keystroke "g" using {command down, shift down}
delay 1
keystroke "/path/to/thumb_[date].jpg"
delay 0.5
keystroke return
delay 1
keystroke return
end tell7. Verify
Take a screenshot of each video's details page to confirm the custom thumbnail is set.
File Size Reference
- Gemini Pro raw output: ~3-4MB (too large)
- After
convert -resize 1280x720 -quality 85: ~200-250KB (well under 2MB limit) - YouTube max thumbnail size: 2MB
- YouTube thumbnail dimensions: 1280x720 (16:9)
- Accepted formats: JPG, PNG, GIF
Example Prompts
Tech Demo Thumbnail
Create a YouTube thumbnail (1280x720) for a video titled "SpaceMolt - AI Agents in Multiplayer Space Games".
This is about a demo of SpaceMolt, a multiplayer space trading and mining game where AI agents autonomously play alongside human players.
The visual should be exciting and space-themed: spaceships, asteroids, neon glow, a dark cosmic background.
Bold text overlay: "SPACEMOLT" prominently, with smaller text "AI AGENTS IN SPACE GAMES" below it.
Include a bottom bar with "AI IN ACTION 7 FEB 2026" in bold.
Please include the Latent Space podcast logo in a corner — search the web for "Latent Space podcast logo".
Use vibrant sci-fi colors — electric blue, green, purple neon tones.Academic Paper Thumbnail
Create a YouTube thumbnail (1280x720) for a video titled "RL via Self-Distillation (SDPO) Paper Reading".
This is an academic paper reading session about SDPO — Self-Distillation Preference Optimization.
The visual should be academic but engaging: neural network or brain visualization with mathematical formulas.
Bold text overlay: "SDPO" prominently, with smaller text "SELF-DISTILLATION PREFERENCE OPTIMIZATION" below it.
Include a bottom bar with "PAPER CLUB 12 FEB 2026" in bold.
Please include the Latent Space podcast logo in a corner — search the web for "Latent Space podcast logo".
Use deep blue, orange, and white color palette with a dark academic-themed background.Troubleshooting
- "File is bigger than 2MB": Compress with ImageMagick
convert -resize 1280x720 -quality 85 - Gemini generation takes long: Pro mode takes 30-45 seconds. Wait and check with a screenshot.
- Upload button not visible: Use the
findtool to locate the hidden "Upload file" button by ref. - File picker automation fails: Instruct user to manually select the file.
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
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.

