web-technique-to-skill

Extracts a working web interaction mechanism into a reusable skill with a verified demo.

5.7k|685|Updated Feb 3, 2026
One-click install
npx skills add https://github.com/MengTo/Skills --skill web-technique-to-skill
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: web-technique-to-skill
Source: https://github.com/MengTo/Skills/tree/main/agent-skills/codex/web-technique-to-skill
Command: npx skills add https://github.com/MengTo/Skills --skill web-technique-to-skill

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

When a page, canvas scene, shader, scroll effect, or hover interaction turns out well, the knowledge of why it works stays trapped in one project's code. This Skill turns that working code into a reusable, versioned skill by isolating the single mechanism that makes the effect work, separating it from project-specific staging, and packaging it with a demo that proves both the mechanism and the visual fidelity.

Core Features & Use Cases

  • Mechanism isolation: Sorts source code into mechanism, staging, and incidental piles, keeping only the mechanism in the skill body while preserving approved reference staging in the demo.
  • Quality contract enforcement: Requires failure-anchored rules, real numeric constants instead of adjectives, symptom-first gotchas, stack disclosure (runtime, renderer, technique libraries), and declared boundaries against neighboring skills.
  • Browser verification workflow: Mandates driving the demo at 1440x900 and 390x844, testing reduced-motion, keyboard focus, multi-state traversal, and a clean console before claiming the technique works.
  • Use Case: After shipping a WebGL night-harbour page with a volumetric lighthouse beam, use this Skill to extract the ray-marched fog density technique into a standalone volumetric-beam-sweep skill with a self-contained demo, leaving the client brand, licensed fonts, and page layout behind.

Quick Start

Use the web-technique-to-skill skill to extract the mechanism behind this working effect and package it as a reusable skill with a verified demo.

Frequently Asked Questions about web-technique-to-skill

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I turn a working web effect into a reusable skill?

Start from working code and write one sentence naming the mechanism that, if removed, makes the effect stop working. Sort every part into mechanism, staging, or incidental piles, keep only the mechanism in the skill body, and build a demo that reproduces the approved reference staging around it.

What is the difference between mechanism and staging in a web technique?

The mechanism is the maths, state model, ordering constraint, or budget that makes the effect work; staging is the palette, copy, imagery, and layout. Test by changing subject, palette, and layout mentally: if the sentence still holds, it is the mechanism.

When should I use this instead of extracting from an article or prompt?

Use this Skill when you built the thing yourself and it works, so the knowledge is trapped in one file. When the source is an article or prompt pack describing behavior rather than working code, reach for the article-prompts-to-skills skill instead.

How do I verify a web technique demo before publishing it?

Drive the demo in a browser at 1440x900 and 390x844, exercise the primary interaction, traverse every state forward and backward for scroll or timeline techniques, tab through for visible focus, run the reduced-motion path, and confirm a clean console at both sizes.

What accessibility requirements apply to extracted web effects?

Under prefers-reduced-motion, render a designed still frame rather than hiding the effect, and keep controls live. Pause on document.hidden, clamp dt to about 1/30s, cap device pixel ratio at 2, size from a ResizeObserver, and keep controls as keyboard-reachable form elements.

Why should skill rules name the failure they prevent?

A rule with a named failure is testable, while a rule without one is decoration. Stating the wrong result, such as an independent sine reading as a wobble instead of a tumble, lets readers recognize the bug they are currently staring at.