title-card
Skill0 starsChanged 8 days agoMITView on GitHub
Make static title cards and thumbnails from HTML templates rendered to PNG, using the same deterministic render path as the html-video skill (frame 0 of the animation is the card). Use when the user wants a thumbnail, a title card, a cover image for a video or post, or says "buatkan thumbnail", "title card", "gambar judul", "cover". Do NOT use for moving content or for editing photographs.
In this repository15 other files it gives its agents.
- cloudflare-workerSkill
- oss-contributorSkill
- small-briefsSkill
- councilSkill
- repo-auditSkill
- vps-deploySkill
- interactive-figureSkill
- lab-reportSkill
- net-labSkill
- study-notesSkill
- vault-routerSkill
- html-videoSkill
- manim-explainerSkill
- caption-craftSkill
- proofread-idSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 45
Verify: view the PNG at full size and at 96px wide.
- Distinctive: line 53
Use the project's real visuals (a frame or diagram from the actual video), never stock decoration that promises content the video does not …
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 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.
Files
Title Card
A good thumbnail is a video's first frame doing a second job. This skill is a spin-off of html-video in this repo: the card is an HTML composition, rendered with the same deterministic capture pipeline, seeking frame 0 instead of capturing a sequence. One source of truth for how a project's visuals look, whether they move or not.
The composition rules
- One message. A title card says one thing: the topic, in the fewest words that stay accurate. If the title needs a comma and an "and", it is two cards.
- Readable at 5% size. The card will mostly be seen as a small rectangle in a feed. Test by shrinking: if the main words cannot be read at thumbnail size, the type is too small or the words too many.
- Two or three elements, no more. Title text, one visual anchor (a diagram piece, a device, a number), and the background. A fourth element competes; remove it.
- Contrast is a feature. Text over a quiet zone of the image, never over the busy part. Dark card, light type, one accent color that matches the video or series.
Templates
assets/ holds three starting points, all 1080x1920 (vertical, for
Shorts-style feeds) unless resized: card-big-type.html (title is
the visual), card-diagram.html (a diagram anchor with the title
under it), card-number.html (one big measured number as the hook).
Copy one, replace text and anchor, keep the layout bones.
Render
- Open the card HTML in the capture pipeline from html-video (Playwright, fixed viewport, device scale 2 for crisp text).
- Capture frame 0 as PNG. No animation runs; if the template inherits motion from a video project, freeze it at the composed state.
- Verify: view the PNG at full size and at 96px wide. Both must work. Check the file is a real PNG of the intended dimensions before delivering.
Rules
- The card must be derivable again: keep the HTML beside the PNG. A thumbnail nobody can regenerate gets remade badly later.
- Use the project's real visuals (a frame or diagram from the actual video), never stock decoration that promises content the video does not have.
- Text on the card is the video's actual title or a faithful shortening of it. Clickbait the content cannot cash is a debt.
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.

