Skip to content
Back to skills

Web

ASecurity

Build and verify interactive HTML experiences in Web Studio's live, sandboxed preview.

  • 1,124 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added September 23, 2026
ai-agentsnodeperformance

Works with

  • cli

Security analysis

A100/100

Pro scans all 3 files and shows the line behind each finding

Scanned September 23, 2026

npx -y skills add autonomous-ai/openharness --skill web --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Web?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Web
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/autonomous-ai-web/badge)](https://www.skillsdirectory.com/skills/autonomous-ai-web)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: web
description: Build and verify interactive HTML experiences in Web Studio's live, sandboxed preview.
---

# Web Studio

`index.html` is the entry point. Local JS, CSS, images and ES modules work through the shared viewer;
use relative URLs. No bundler is required. Prefer local assets for a reliable first open.

Field's starter uses WebGL2 with a Canvas 2D fallback. Its geometry, palette and controls are all
editable. Keep GPU resources and geometry buffers reusable; coalesce continuous updates in rAF.
Pause animation when hidden and respect reduced motion. Controls must still work when paused.
A new visual direction can replace the starter entirely when that serves the user's request.

## Verify the experience

Keep `proof.json` aligned with the current page. Each action names a CSS `selector`, followed by
`click: true`, `fill` or `select`; assertions use exact `text` or
`attribute: {"name":"aria-pressed","value":"true"}`. Include a state change and its meaningful result.

```sh
node "$WEB_SKILLS/web/scripts/screenshot.mjs" "$HARNESS_WORKSPACE/index.html"
node "$WEB_SKILLS/web/scripts/perf.mjs" "$HARNESS_WORKSPACE/index.html"
```

The proof runs the real sandboxed HTTP viewer, checks the actions and browser errors, saves
`.harness/last.png` and a dependency-hashed `browser-proof.json`, and refreshes the verdict.
Inspect the image and test a narrow viewport as well as desktop.

The performance probe drives those controls while measuring rAF delays and long tasks, rather than
measuring an idle page. Target ≥55 fps and <2% slow frames. Record the hardware/renderer and dataset;
a software WebGL renderer is not a measurement of the user's GPU. `PLAYWRIGHT_CHANNEL=chrome` can
measure installed Chrome. Do not label an unmeasured or failing page “60 fps.”

Helpers need Playwright plus Chromium, installed in the shared isolated-web-viewer package or addressed by
`PLAYWRIGHT_MODULE` (absolute path to `playwright/index.mjs`). Missing tooling is a failed proof,
not permission to report success. The preview itself needs only Node 20+.

Files in this skill

  • SKILL.md2 KB
  • scripts/perf.mjs713 B
  • scripts/screenshot.mjs714 B

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…