Agent skill · Audited from source
finesse-ui
skill · mouse-lin/finesse-skillMIT licenceAudited 20 Oct 2026 · source 7e538389b569
Is it safe? Yes — SAFE, grade B (89/100). We read this skill's source at commit 7e538389b569 and found no filesystem, network or shell access and no hard-coded secrets. What it does: Build never-cheap, high-craft web interfaces — brand surfaces (landing pages, brand sites, launches, portfolios, hero pages with real WebGL/Three.js/Canvas/GSAP engines), product UI (dashboards, admin panels, analytics,…
Cite
Cite this page
ModelsAtlas. “finesse-ui (mouse-lin/finesse-skill)”. Retrieved 20 Oct 2026. https://modelsatlas.com/skills/mouse-lin/finesse-ui
Trust audit
Audit 0.4.1 · source read 20 Oct 2026
SAFE means nothing in the source contradicts what the skill says it does. Grade B is shared by 13,985 of 14,816 audited skills.
What the source does
- FilesystemNone observedNo reads or writes in the source
- NetworkNone observedNo outbound calls in the source
- ShellNone observedNo commands run in the source
- DependenciesPinnedNothing floats
- SecretsNone foundNo keys in source
Five audit layers
| Layer | What we check | Result | Findings |
|---|---|---|---|
L0 Provenance | Repo owner, licence, archive status, activity | PASS | 0 |
L1 Static source | Secrets, path traversal, supply chain, hidden code | PASS | 0 |
L2 Instructions | Prompt injection, authority framing, hidden directives | PASS | 0 |
L3 Inventory | The files and declarations match the source | PASS | 0 |
L4 Sandbox | Run in isolation and watch behaviour | SKIPPED why | 0 |
Gates applied: no behavioural pass.
Findings (0)
No findings at any severity. We re-audit when the source changes.
Audit 0.4.1 · source sha 7e538389b569 · 20 Oct 2026Report an issue or request a re-scan
Provenance
Who published it, and what the verdict is pinned to
- Publishermouse-lin GitHub userOwner of the repository this skill is read from
- Source foldermouse-lin/finesse-skill/skills/finesse-uiThe exact folder we audited
- LicenceMITAs the repository states it
- Commit pinned for this verdict7e538389b569A new audit runs when the source changes
- Repository statusNot archivedLast push 77 days ago
- Sandboxed runNot runSkills are read, not executed; layer L4 is skipped
What it tells the agent
Full source on GitHub ↗Its description: Build never-cheap, high-craft web interfaces — brand surfaces (landing pages, brand sites, launches, portfolios, hero pages with real WebGL/Three.js/Canvas/GSAP engines), product UI (dashboards, admin panels, analytics,…
--- name: finesse-ui description: 'Build never-cheap, high-craft web interfaces — brand surfaces (landing pages, brand sites, launches, portfolios, hero pages with real WebGL/Three.js/Canvas/GSAP engines), product UI (dashboards, admin panels, analytics, data tables, app shells), workflow UI (merchant/admin consoles, publish & create wizards, config and settings pages, review queues), AI workbenches (AI 工作台, agent consoles, 智能体控制台, copilot UI, run streams, approval/human-in-the-loop cards), commerce pages (product detail pages, listing/category pages, cart, checkout), and H5 / phone-only screens (app UI prototypes, 活动页, mobile PDP, data-report H5, mobile sites). Routes by register: brand → soul + spectacle engine; product → palette + component system + density (split further into pages you READ = dashboards, pages you OPERATE = consoles/wizards, and and a DELEGATION layer that stacks on either when an agent does the work); commerce → PDP/PLP skeletons + anti-dark-pattern rules; h5 → a fixed 390×844 frame, safe-area math, thumb-zone hierarchy, and native furniture (status bar, TabBar, bottom sheet, FAB, push & FLIP transitions) across six phone morphologies. Ships a product color library (tinted neutral ramps + 16 accents + 12 paste-ready sets) so dashboards stop defaulting to blue. Always reads the brief first and audits against an anti-slop cheapness blacklist. Ships an anti-sameness layer (divergence) that composes a soul from five orthogonal axes instead of picking from a persona list, so repeat runs stop converging on one look. Routes single-element briefs to a component flow (8 mandatory states + a preview file) instead of the page apparatus. Carries a cross-run build log so rotation is enforced rather than narrated, and a mobile floor covering the six ways a crafted page breaks on a phone. Supports verb commands (audit · bolder · quieter · soul · diverge · animate · densify · redesign) for targeted iteration. Ships a motion layer that separates WHAT moves from HOW it is built: six routes (CSS · native scroll-driven · View Transitions/WAAPI · GSAP · Canvas/WebGL · zero-dependency CSS 3D space) against a ten-family effect catalogue of ~80 variants (spatial camera moves, particle fields, fluid & material, scroll narrative, typographic, image transform, geometric construction, physics & inertia, state transition, atmosphere), each family carrying its slop form; an effect→route lookup so a page stops shipping 60KB to do four lines of work; and a four-beat sheet that replaces "one spectacular hero over generic fade-ups" — because 「炫酷」 and 「好看」 select nothing, and every model answers them with the same particle hero. Separates 后台 (revolves around a batch of business objects) from 工作台 (revolves around one thing he does over and over) — one fork that decides density, soul and shell, resolved by what the thing revolves around rather than by head-count or screen; a workbench then picks a body, desktop or phone morph A.1. Ships an AI-workbench layer that stacks on any of them for the pages agents actually run in — a run stream instead of a chart, nine run states, a resident stop control, in-stream approval cards with linked evidence, and cost as a first-class receipt. Triggers on "make this look premium", "landing page", "personal site", "portfolio", "个人主页", "落地页", "lookbook", "dashboard", "admin panel", "商家后台", "工作台", "专属工作台", "个人工作台", "每日工作台", "打卡页", "记录页", "daily desk", "back-office", "console", "publish flow", "发布流程", "wizard", "settings page", "analytics UI", "data table", "app UI", "product page", "PDP", "listing page", "checkout", "dashboard colors", "AI 工作台", "agent console", "智能体控制台", "AI 助手界面", "copilot UI", "运行记录", "工具调用", "人工审批", "值守台", "H5", "H5 页面", "移动端页面", "手机端页面", "活动页", "小程序页面", "公众号页面", "app 原型", "mobile app UI", "app screen", "in-app page", "移动端商详", "报告 H5", "give it a soul / a vibe", "anti-slop", "hero animation", "动效", "炫酷", "加点动画", "让它动起来", "太静了", "滚动动画", "页面转场", "沉浸式", "走进去", "scroll animation", "page transition", "make it move", "feels generic", "every page looks the same", "每次都差不多", "make me a button", "just this component", "按钮", "单个组件", "component states", "breaks on mobile", "手机端错位", "/finesse".' version: 0.20.0 user-invocable: true argument-hint: "[craft · audit · bolder|quieter|soul|diverge · animate|depth|densify · redesign · init|document] [target]" license: MIT metadata: short-description: Craft interfaces that never look cheap --- # finesse — Technically Spectacular · Soul-Distinct · Never Cheap > **finesse builds two kinds of interface and routes by register (§0):** > - **brand** — design IS the product: landing pages, brand sites, launches, portfolios, hero pages. Optimize for **spectacle + soul + first impression** — a real visual engine, an opinionated personality. > - **product** — design SERVES the product: dashboards, admin panels, analytics, data tables, app shells, settings. Optimize for **clarity + density + usability** — and still never cheap. > > The through-line is identical: **high craft, zero AI-slop.** What applies to **both** is the **universal craft floor** (tinted neutrals, no `#fff`/`#000`, translucent/hairline borders, tinted shadows, contrast floors), the cheapness blacklist (§6), and the pre-flight (§8). What **forks** is the *substrate above that floor* and the middle: **brand** lays the §3 brand substrate (`design-dna.md` — grain, vignette, display type) and reaches for a hero engine (§4); **product** lays its **own** substrate (`product-ui.md` §0 — surfaces, cards, KPI tiles, density) and reaches for a component system + data viz. A dashboard is a **different design language**, not a brand page with charts — it does **not** inherit grain / vignette / giant hero type / dark-default / a hero engine. > > **Two more registers sit alongside them, and they work differently from each other.** **commerce** (PDP/PLP/cart/checkout) is a *hybrid* — route it to brand or product by what the specific page is doing. **h5** (phone-only screens:
Excerpt, first 6,000 characters of SKILL.md, verbatim from the audited commit. It is text from someone else’s repository, quoted so you can judge it before your agent reads it.
Install
Pinned to the audited commit 7e53838
git clone https://github.com/mouse-lin/finesse-skill git -C finesse-skill checkout 7e538389b569
The author documents no install command, so this uses the repository directly. Our verdict covers commit 7e538389b569 only.
Works in which agents?
What its own text says, not a test
| Agent | Status | More |
|---|---|---|
| Claude Code | Mentioned in its text, not tested | Skills for Claude Code · About Claude Code |
| Codex | Mentioned in its text, not tested | Skills for Codex · About Codex |
| GitHub Copilot | Mentioned in its text, not tested | Skills for GitHub Copilot · About GitHub Copilot |
| Cursor | Mentioned in its text, not tested | Skills for Cursor · About Cursor |
Audit history
Every audit, every source change
- ✓SAFE B 89 / 100
first audit · source
7e538389b569
Similar skills
All design & front-end skills →obra/superpowers
Use when about to claim work is complete, fixed, or passing, before committing or creating PRs - requires running verification commands and confirming output before making any success claims; evidence before assertions…
obra/superpowers
Use when you have a spec or requirements for a multi-step task, before touching code
obra/superpowers
Use when starting any conversation - establishes how to find and use skills, requiring skill invocation before ANY response including clarifying questions
Questions about finesse-ui
Is finesse-ui safe to install?
What can finesse-ui access on my machine?
Which assistants does finesse-ui work with?
How current is this page?
Sources
- github.com/mouse-lin/finesse-skill/tree/main/skills/finesse-ui — the audited source at commit 7e538389b569 and every finding aboveObserved · 20 Oct 2026
Spotted something wrong? Report a correction · Methodology
Keep going in your AI
Opens with this page as the source. We add nothing else to the prompt.
- Is finesse-ui safe to install?
- What can finesse-ui access on my machine?
- Which assistants does finesse-ui work with?