How to use the VFX Text Cursor agent skill

What the VFX Text Cursor skill makes the agent do

Once installed, the agent can turn a quote you supply into a single-file HTML page. The text is centered on a dark background with a subtle vignette, and each character appears with a short delay (80 ms per character in the recipe). A cursor follows the reveal point, and a motion-blur trail and a few angled light leaks sit behind it.

When the line is finished, a shimmer band sweeps across the text. Revealed letters are white, and the reveal moment briefly shows pink and cyan offset shadows that settle back to normal within about 200 ms.

This is a visual style recipe, not a video editor. The output is a web page. You record it or capture its screen yourself.

• Canvas: 1920×1080 (16:9)
• Fonts referenced by name: Inter Tight, Source Sans 3, Noto Sans SC
• Caption at top (for example FRAME 01 · OPENING), subtitle line below the quote, timecode at bottom right
• Accent colors: hot pink, cyan or amber for the cursor

Install the skill from the ZIP download

Download the ZIP for vfx-text-cursor from our page and unzip it. The package contains SKILL.md, example.html and example.md.

For a personal install, place the unzipped skill folder in ~/.claude/skills/. To limit it to one project, place it in that project's .claude/skills/ folder instead.

If you use the Claude apps instead of files on disk, you can upload the skill under Settings → Capabilities → Skills.

• Personal: ~/.claude/skills/vfx-text-cursor/
• Project only: .claude/skills/vfx-text-cursor/ inside your project
• Claude apps: Settings → Capabilities → Skills

Trigger the skill with a natural request

You do not need a special command. Describe what you want in plain language and mention the quote and the video use. The agent should pick up the skill from its description when the request matches a title-card or text-reveal task.

Always supply the exact quote. The skill is written to use only the text you provide and not to invent wording.

• "Make a 1920×1080 opening title screen for my video with this quote, typing letter by letter with a glowing cursor: [your quote]"
• "Build an HTML hero frame with a chromatic text reveal for this line, using hot pink and cyan."
• "Create a video intro card that reveals this quote with a shimmer sweep when it finishes."

Examples of what you can ask for

The bundled example shows a Chinese-language quote with a markdown source file, a FRAME 01 · OPENING caption and a timecode of 00:03:21. You can use it as a reference for layout and tone.

You can adapt the same recipe for other quote reveals by changing the text, the accent color (hot pink, cyan or amber) or the caption. Keep the effects to one accent pair so the result stays clean.

• Keep the bundled example as a starting point and swap in your own quote
• Change the accent to amber for a warmer look
• Ask for a caption and subtitle that match your chapter name

Safety notes: scripts and network

The skill does not run scripts and does not use the network. Its output is a single HTML file that references only the named fonts. No API keys or accounts are needed.

The skill does offer a reduced-motion option, which shows all the text at once. Still, flashing and chromatic effects can be hard for some viewers, so check the result before you publish it.

• Runs scripts: no
• Uses network: no
• Requires a web browser to view the HTML file
• Review the generated HTML before sharing it

Tips for a better result

Give the agent the exact final wording and the target length of the video opener. Ask for the reduced-motion version too if the video will be shown widely. Check the fonts on your machine, since the page refers to Inter Tight, Source Sans 3 and Noto Sans SC by name and may fall back to other fonts if they are not installed.

• Provide the quote in its final form before generating
• Ask for the reduced-motion version as well
• Install the named fonts so the page looks as intended
• Record the page at 1920×1080 with your own screen recorder