agentleFS
imMamdouhaboammarimMamdouhaboammar/Mimera/.agents/skills/website-cloner/SKILL.md

website-cloner

Skill1 starsChanged 33 days agoMITView on GitHub

Agentic workflow for cloning websites with pixel-perfect fidelity using Mimera's visual simulation engine and specialized sub-agents. Use when cloning/reproducing a website, extracting design DNA, or rebuilding an interface with React, Tailwind CSS, and Motion. Enforces Bun runtime, zero-guessing computed style extraction, dynamic video normalization, and the strict +90% visual parity threshold gate.

5 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

2 rules. 0 are common, 2 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 4 edited versions we read keep these.

  • 2

    Distinctive

    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
  1. Distinctive:

    Never guess padding, margins, font sizes, or line heights.

    line 64
  2. Distinctive:

    Always extract exact values from the live DOM via Playwright:

    line 65

“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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Website Cloner Skill (Powered by Mimera Engine)

Clone any website with pixel-perfect fidelity using an orchestrated multi-agent workflow integrated with Mimera's native visual simulation and verification engine.

Overview

This skill integrates the multi-agent cloning pattern with Mimera's zero-dependency visual engine:

  • Orchestration: End-to-end management from reference reconnaissance to verified component locking.
  • 4 Specialized Sub-Agents:
    1. website-screenshotter: Captures desktop (1440x900) and mobile (390x844) side-by-side viewports at 700px scroll intervals.
    2. website-extractor: Harvests assets to public/, extracts live computed CSS (getComputedStyle), fonts, and color palettes.
    3. website-cloner: Implements responsive React components with Tailwind CSS and Motion.
    4. website-qa-reviewer: Evaluates structural edge MSE, color histograms, and pixel diffs via bun mimera simulate.
  • Quality Gate: Non-negotiable +90.0% minimum visual parity threshold (target 100%).
  • Runtime Policy: Bun is strictly mandatory. npm and yarn are prohibited.

Architecture

┌───────────────────────────────────────────────┐
│           ORCHESTRATOR / MIMERA CLI           │
│     bun mimera prepare -> capture -> analyze  │
└───────────────────────────────────────────────┘
                       │
       ┌───────────────┴───────────────┐
       ▼                               ▼
┌──────────────┐              ┌────────────────┐
│ screenshotter│              │   extractor    │
│ (Dual view)  │              │ (Computed CSS) │
└──────────────┘              └────────────────┘
       │                               │
       └───────────────┬───────────────┘
                       ▼
              ┌────────────────┐
              │     cloner     │◄──────────────┐
              │(React+Tailwind)│               │
              └────────────────┘               │
                       │                       │
                       ▼                       │
              ┌────────────────┐               │
              │  qa-reviewer   │───────────────┘
              │(Parity >= 90%) │  (Loop until >= 90.0% or 5 iterations)
              └────────────────┘
                       │ (Pass >= 90%)
                       ▼
              ┌────────────────┐
              │ regression-lock│
              │(Cryptographic) │
              └────────────────┘

Golden Principles

  1. Zero-Guessing Computed Style Extraction:
    • Never guess padding, margins, font sizes, or line heights.
    • Always extract exact values from the live DOM via Playwright:
      const computed = window.getComputedStyle(element);
      console.log(computed.fontSize, computed.lineHeight, computed.padding, computed.margin);
  2. Dynamic Noise Normalization:
    • Normalize async video frames, YouTube embeds, and dynamic canvases before comparison:
      document.querySelectorAll('iframe, video, canvas').forEach(el => el.style.opacity = '0.3');
  3. Multi-Viewport Side-by-Side Verification:
    • Evaluate both Desktop (1440x900) and Mobile (390x844) at 700px vertical increments.
  4. Strict Parity Gate (+90% Minimum):
    • Parity combines Structural Edge MSE (65%), Pixel Diff (25%), and Color Histogram (10%).
    • Score < 90.0% automatically rejects the build.
  5. Regression Locking:
    • Once approved, seal all modified files using bun mimera lock -c <component-id> -f <files...>.

Step-by-Step Workflow

Step 1: Initialize Project & Reconnaissance

# Initialize session with reference URL
bun mimera init --reference <target-url> --host gemini-cli --mode high-fidelity

# Capture multi-viewport evidence pack and download assets into public/
bun mimera capture

Step 2: Extract Design DNA & Tokens

# Extract typography hierarchy, rhythm unit, and color tokens
bun mimera analyze

Step 3: Implement Component

  • Build the component in src/components/ or app/ using React, Tailwind CSS arbitrary values (bg-[#hex], p-[18px]), and Motion.
  • Connect self-hosted fonts in index.html or root CSS.

Step 4: Run Visual Simulation & Parity Audit

# Run automated side-by-side dual-viewport audit
bun mimera simulate --reference <target-url> --target http://localhost:5173/ --threshold 90

Step 5: Review & Seal Regression Lock

# Multi-reviewer inspection
bun mimera review --component <component-id>

# Seal approved code
bun mimera lock --component <component-id> -f src/components/<Component>.tsx

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.