UI-perfect CLAUDE.md

UI-perfect CLAUDE.md is an instructions file for Claude Code from nv-minh/UI-perfect. It costs 1,139 tokens per session, scanned A, original, MIT.

Repository-specific instructions for Claude Code in UI-perfect, a tool that checks how closely a website implementation matches a Figma design. Figma is a design and prototyping tool; the project compares design measurements with browser measurements and uses pixel differences for individual elements.

In plain words
What is it for?
Use them when working on UI-perfect or Specgate to inspect Figma comparisons, run verification and audit commands, configure test screens, and maintain the core, Figma, browser-probe, and CLI packages.
Why use it?
They explain the repository structure, verification commands, configuration files, and rules for measuring design accuracy. This helps the assistant diagnose mismatches using measurements and the correct project components.

Instructions file for Claude Code

Written for Claude Code: the file is CLAUDE.md. Also seen: reads .claude/ paths; mentions Claude Code.

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add instructions/nv-minh/ui-perfect/claude-md
Clone the repo
git clone --depth 1 https://github.com/nv-minh/UI-perfect

Made for: Claude Code.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for UI-perfect CLAUDE.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/nv-minh/ui-perfect/claude-md.svg)](https://agentmods.dev/instructions/nv-minh/ui-perfect/claude-md)
Your own site
<a href="https://agentmods.dev/instructions/nv-minh/ui-perfect/claude-md"><img src="https://agentmods.dev/badge/instructions/nv-minh/ui-perfect/claude-md.svg" alt="Measured on agentmods" height="20"></a>
Per session 1,139 This file is loaded in full into every session.
When invoked 1,139 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.01139 $0.01139
Opus 5 $0.00570 $0.00570
Sonnet 5 $0.00228 $0.00228
Haiku 4.5 $0.00114 $0.00114

Measured 5d ago against content hash 79aaecb75767, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-06, from the pricing page.

Security

Grade A, and why

UI-perfect CLAUDE.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 5d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

CLAUDE.md · 54 lines

How it starts

The opening of the file, as written. The whole thing — 54 lines — stays where its author put it; the contents beside it link to each section on GitHub.

UI-perfect / specgate

Tool verify độ khớp Figma design ↔ implementation cho Claude Code. Measurement-first: so số đo (Figma REST vs getComputedStyle) là nguồn sự thật, pixel diff chỉ per-node để bắt sai asset. Design docs: PLAN.md, PLAN-02-lockfile-mobile.md. Plan thực hiện đã duyệt: ~/.claude/plans/t-i-c-t-o-repo-deep-sky.md.

Cấu trúc

  • packages/core — NormalizedNode types, Probe interface, compare engine. KHÔNG import Figma/Playwright.
  • packages/figma — REST client (batch + cache theo version + 429 backoff) + normalize.
  • packages/probe-web — Playwright WebProbe + freeze kit.
  • packages/clispecgate verify|accept|guard|audit|impact|mcp (6 lệnh).
  • apps/testbed — checkout screen dựng theo Figma (baseline hiện tại: 0 HIGH / 0 MED / 1 LOW, coverage 38% node significant đã map) + mutations/ (harness 3 bộ detect/benign/mapping).
  • .specgate/screens.config.json — fileKey, figma.timeout, app (url/command/fixtureMode), sources (cho impact), impact.always, và logical key map per viewport.
  • .research/ — repo tham chiếu (gitignored): uimatch, looks-same, Figma-Context-MCP, rest-api-spec.

Quy tắc quan trọng

  • Playwright pin exact 1.62.1 cả trong packages/probe-web/package.json lẫn docker/docker-compose.yml (image + npx). Lệch version = lỗi 428 khi connect.
  • Browser chạy trong Docker: cd docker && docker compose up -dws://localhost:6003/specgate. Testbed Vite phải giữ host: true để container thấy http://host.docker.internal:5173.
  • Figma Tier-1 rate limit ~10 req/min (Starter plan): mọi code gọi Figma phải cache theo (fileKey, version) và backoff khi 429. Không bao giờ gọi Figma trong vòng lặp không throttle.
  • FIGMA_TOKEN chỉ nằm trong .env (gitignored). CLI cũng chấp nhận FIGMA_ACCESS_TOKEN làm fallback (tên biến của uimatch) — có cả hai thì FIGMA_TOKEN thắng. .env nhận cả dạng export KEY=value.
  • Mọi request Figma chạy dưới 1 deadline duy nhất (figma.timeout trong config, mặc định 30s) bao trọn cả header lẫn body — kể cả lúc tải PNG từ S3. Hết giờ → Figma request timeout sau 30000ms tại <endpoint>, đó là lỗi hạ tầng KHÔNG phải lỗi UI.
  • DPR đọc từ config (viewport.dpr, mặc định 2) và phải khớp nhau ở cả 3 nơi: viewport probe, Figma export scale, screenshot. Testbed dùng 2 — đổi thì đổi một chỗ trong screens.config.json, đừng hardcode rải rác.
  • Khi generate UI từ Figma: mỗi element ứng với 1 Figma node phải có data-fig="<logical key>" (key trong screens.config.json, KHÔNG phải node id). Root frame thêm data-fig-root.

Read the full file on GitHub · 54 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 5d ago First seen · 54 lines · 1,139 tokens per session scan A 79aaecb75767

Subscribe to this mod's changes

UI-perfect CLAUDE.md is an instructions file published in the GitHub repository nv-minh/UI-perfect (0 stars, last pushed 1mo ago), licensed MIT. It adds 1,139 tokens to every session, about $0.0057 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.