Agent skill

Nova Frontend

by iZiTTMarvin in iZiTTMarvin/Nova-agent

Anti-slop, high-craft frontend design: landing pages, brand sites, portfolios, dashboards, product/admin UI, commerce (PDP/PLP/checkout).

MITAuto-check passedFrontend & Design

Install Nova Frontend

skills CLI
$ npx skills add iZiTTMarvin/Nova-agent --skill nova-frontend -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install iZiTTMarvin/Nova-agent nova-frontend --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/iZiTTMarvin/Nova-agent.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.nova/skills/nova-frontend .claude/skills/nova-frontend && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
nova-frontend
GitHub stars
178
Token cost
~1.5k tokens
SKILL.md length
450 words
Files
11 (incl. references)
Skills in repo
7
Repo updated
First seen
Licence
MIT

At a glance

Anti-slop, high-craft frontend design: landing pages, brand sites, portfolios, dashboards, product/admin UI, commerce (PDP/PLP/checkout).

  • Works in 6 steps: §0 读懂 brief:判定 register(brand / product… → §1 定三个 dial:SOUL / SPECTACLE / DENSITY,从… → 两条路都要:铺 §3 的 premium substrate(读… → …
  • Premium / landing page / dashboard / redesign / give it a soul / /nova-frontend
  • SKILL.md covers 如何使用(执行顺序), §0 读懂 Brief(先于一切), §1 三个 Dial and 常驻底线(即使不读任何 reference 也必须遵守), plus 2 more sections
  • Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md

What it does

Nova Frontend is an agent skill from iZiTTMarvin/Nova-agent. Anti-slop, high-craft frontend design: landing pages, brand sites, portfolios, dashboards, product/admin UI, commerce (PDP/PLP/checkout). Reads the brief, sets register + dials, loads references on demand. Never looks AI-templated. Triggers on premium / landing page / dashboard / redesign / give it a soul / /nova-frontend.

Its SKILL.md is about 1.5k tokens, which your agent loads only when the skill is triggered. The skill folder holds 11 other files, including reference files (for example `references/anti-cheap.md`, `references/commerce-ui.md` and `references/design-dna.md`).

It sits in Frontend & Design, covering Landing pages and Humanizing AI text. The repository describes itself as: 最适合新手、懒人、业余开发者使用的本地桌面 Agent,长会话不丢上下文,一句话跑完全流程,无需如何门槛即可完成高质量的开发. The licence is MIT.

When your agent uses it

  • Premium / landing page / dashboard / redesign / give it a soul / /nova-frontend
  • Tasks that involve Landing pages
  • Tasks that involve Humanizing AI text

Example prompts

  • “/nova-frontend”

Workflow steps

6 steps, taken from the first numbered list in SKILL.md.

  1. §0 读懂 brief:判定 register(brand / product / commerce)与灵魂方向,输出一行 Design Read。
  2. §1 定三个 dial:SOUL / SPECTACLE / DENSITY,从 brief 推断,不要默默用基线。
  3. 两条路都要:铺 §3 的 premium substrate(读 design-dna.md)——这层物理质感让 dashboard 也不廉价。
  4. 分流
  5. 组装页面,动效必须被动机驱动。
  6. 发布前跑 §6 反廉价扫描(anti-cheap.md)+ 完整 pre-flight(preflight.md)。

What it can do on your machine

Read from SKILL.md and the folder at commit 216466e. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    No scripts in the folder and no shell commands in SKILL.md.

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Nova Frontend loads about 1.5k tokens when it runs, and up to ~26k if it reads all its reference files. Until then it costs about 85 tokens; SKILL.md has 450 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~85
When it runs · the whole SKILL.md, loaded when a task matches
~1.5k
With references · SKILL.md plus every file in references/, read only if the agent opens them
~26k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from iZiTTMarvin/Nova-agent at commit 216466e, republished under its MIT licence (© iZiTTMarvin). 450 words, ~1,519 tokens.

Download SKILL.mdSave it as .claude/skills/nova-frontend/SKILL.md (or your agent's skills folder). This skill also uses 10 other files; get the full folder from GitHub.
name
nova-frontend
description
Anti-slop, high-craft frontend design: landing pages, brand sites, portfolios, dashboards, product/admin UI, commerce (PDP/PLP/checkout). Reads the brief, sets register + dials, loads references on demand. Never looks AI-templated. Triggers on premium / landing page / dashboard / redesign / give it a soul / /nova-frontend.
name_zh
前端高质量设计
argument-hint
[craft|audit|redesign|bolder|quieter|soul|animate|densify] [目标]
user-invocable
true

nova-frontend — 技术精湛 · 灵魂鲜明 · 从不廉价

一个把"防 AI slop"做到极致的前端设计技能。融合三套生产级方法论:先读懂 brief, 定 register 与 dial,再按需加载深层 reference。目标是产出一眼看不出是 AI 套模板的界面。

核心机制:本技能的深层规则拆分在 <%= skillDirectory %>/references/ 下。 SKILL.md 只是路由表 + 常驻底线。每个阶段只用 read 工具读取当下需要的那一个 reference 文件, 不要一次性全读进来。所有规则都是上下文相关的——先读 brief,再挑适用的拉进来。 一套方法对所有 brief 产出同一个页面,就是失败。


如何使用(执行顺序)

  1. §0 读懂 brief:判定 register(brand / product / commerce)与灵魂方向,输出一行 Design Read。
  2. §1 定三个 dial:SOUL / SPECTACLE / DENSITY,从 brief 推断,不要默默用基线。
  3. 两条路都要:铺 §3 的 premium substrate(读 design-dna.md)——这层物理质感让 dashboard 也不廉价。
  4. 分流:
    • brand → 挑一个灵魂(style-personas.md)+ 造一个 hero 引擎(motion-engines.md)。
    • product → 组件系统 + 数据密度(product-ui.md)。
    • commerce → PDP/PLP 骨架 + 反暗黑模式(commerce-ui.md)。
  5. 组装页面,动效必须被动机驱动。
  6. 发布前跑 §6 反廉价扫描(anti-cheap.md)+ 完整 pre-flight(preflight.md)。
Reference 路由表(用 read 按需加载,路径前缀 <%= skillDirectory %>/references/)
何时加载文件
铺 premium substrate(grain / 排版张力 / OKLCH 色彩 / palette+字体表)design-dna.md
brief 指向真实设计系统(Fluent/Carbon/Material/shadcn/GOV.UK…)——附安装命令与官方文档design-systems.md
造 hero 引擎、GSAP 滚动骨架、动效词汇、reduced-motion、框架集成motion-engines.md
dashboard / admin / 数据表 / 图表 / 表单 / 交互状态(product register)product-ui.md
商品详情 PDP / 列表 PLP / 购物车 / 结账 + 反暗黑模式(commerce register)commerce-ui.md
挑灵魂 / persona,或选字流程与 reflex-reject 字体表style-personas.md
需要模式名词汇(hero/nav/scroll/gallery 等)或 Apple Liquid Glass 诚实近似patterns.md
改造已有页面(audit-first,不是推倒重来)redesign.md
完整反廉价黑名单(发布前必扫)anti-cheap.md
最终发布前检查清单preflight.md

§0 读懂 Brief(先于一切)

多数 AI 设计之所以差,是因为模型跳过读题、直接套一个默认审美。别这样。

0.A 判定 register(这一步分流后面所有决策)
  • brand — 设计就是产品:landing page、品牌站、发布页、作品集、hero 页。要大胆、有观点、有奇观。走灵魂 + hero 引擎路线。
  • product — 设计服务于产品:dashboard、admin、分析、数据表、app 外壳、设置。为清晰 / 密度 / 可用性优化。走组件系统路线。仍然从不廉价(继承 substrate + 反廉价黑名单)。
  • commerce — 混合态:按具体页面的任务分流:卖单个 hero 商品的 PDP 偏 brand(挑灵魂但保持规格/评价/信任信号的高密度);多 SKU 的 PLP/市场偏 product(高密度、低 spectacle)。拿不准就问一句:"这页是卖一个产品的氛围,还是帮人比较筛选很多个?"

先读项目记忆:若项目根有 PRODUCT.md / DESIGN.md,先读,它覆盖你的猜测。已有品牌色/字体时——身份保留优先,reflex-reject 列表只对全新决策生效。

0.B 生成一行 Design Read(写代码前必出)

格式:Design Read: {行业} · {2-3词灵魂} · register={brand|product|commerce} · SPECTACLE={n} · hero-engine={类型}

例:Design Read: 深空天文 · 电影感+庄严 · register=brand · SPECTACLE=8 · hero-engine=Three.js 粒子星系

输出 Design Read 后先停,等用户确认方向或纠偏,再进 §1。 除非用户已明确说"直接做"。

0.C brief 模糊时,只问一个问题,别瞎猜

一个尖锐的问题胜过五轮错误默认。问那个最能改变产出的:"这该偏克制编辑感,还是极致奇观?" / "离开 10 秒后你希望访客记住什么?"问完等答案。

0.D Anti-Default 纪律(最常被考的 AI tell)

先说出这个 brief 的懒惰默认,再打败它:"咖啡品牌 → 默认是暖米色 + 黄铜衬线。我拒绝它,改用 {x},因为 {理由}。" 两个高度都要查:一阶(能从行业本身猜出主题+配色吗?"AI SaaS → 紫色辉光 + Inter");二阶(能从行业+你的反参照猜出吗?"不做 SaaS 米色的 AI 工具 → 编辑体排版")。两个答案都不明显才算过。完整 reflex 列表见 anti-cheap.md §0。


§1 三个 Dial

从 Design Read 明确设定,别默默用基线。它们驱动下游一切。

Dial1–34–67–10
SOUL — 个性/品牌化程度中性、安全、系统默认有清晰 vibe独一无二、不可复制的身份
SPECTACLE — 视觉引擎的技术野心(本技能的招牌 dial)纯静态 + CSSGSAP 滚动、Canvas 2D 点缀Three.js/GLSL/WebGL-FBO hero、生成式、滚动钉住的电影感
DENSITY — 每屏信息量通透、一屏一个想法平衡编辑级、数据密集

快速预设(有更强信号立刻覆盖):landing(无其他线索) SOUL7/SPEC6/DEN4 · agency/creative landing 8/8/3 · 高端消费 8/5/3 · portfolio 8/6/3 · editorial 8/4/6 · tech/AI/SaaS 营销 6/7/5 · B2B/企业 4/3/6 · dashboard/analytics(product)4/2/9 · PDP 6/4/6 · PLP 3/2/8 · public-sector/信任优先 3/2/5。

"SPECTACLE 声称即兑现"(强制):若 SPECTACLE ≥ 7,页面必须真有一个能跑的视觉引擎(真实 Three.js/Canvas/GLSL/滚动钉住),能优雅降级,中端设备维持 60fps。声称 8 却发一个渐变色块 = 破损。做不出就把 dial 降到 4,发一个打磨到位的静态页。永远别半吊子做个会卡/断裂的引擎。


Show full SKILL.md (148 more words)Show less

常驻底线(即使不读任何 reference 也必须遵守)

这些是三套方法论里最高频、最致命的 tell。下面每一条都是发布阻断项,完整清单在 anti-cheap.md:

  1. Em-dash(—)完全禁用——标题、eyebrow、pill、正文、引用、署名、按钮、alt 文本,一个都不许有。这是 #1 AI tell。用句号、逗号或重构句子。en-dash(–)当分隔符同禁;范围用连字符(2018-2026、€40-80k)。
  2. 禁 div 假截图——不用 <div> 拼假 dashboard/假任务列表/假终端。用真截图、生成图、真组件预览,或干脆不放。
  3. 禁渐变文字(background-clip:text + gradient)作默认花活。用单色,靠字重/字号强调。
  4. 禁 AI 紫/蓝辉光——紫色渐变按钮、霓虹网格背景、发光 CTA 作默认。用中性底 + 一个高对比强调色。#6366f1(通用靛蓝)是 AI UI 最被滥用的 hex,别默认伸手。
  5. 禁假精确数字——92%、4.1×、5.8mm 之类为凑规格感编造的。要么有真实来源,要么标注 mock,要么删掉。
  6. eyebrow 限量——标题上方那个小号大写宽字距标签,每 3 个 section 最多 1 个(hero 算 1 个)。机械检查:数 uppercase tracking 小标签数量,超过 ceil(sectionCount/3) 就失败。
  7. 禁纯 #fff/#000、硬 #333 边框、亮底上的纯黑投影。用向品牌色微调的中性色、半透明边框、带色相的投影。
  8. 一个强调色锁全页——暖灰站不能在第 7 节冒出蓝色 CTA。选定即锁死,发布前逐组件审。
  9. 禁默认品类配色——craft/heritage 的米色+黄铜、AI/SaaS 的紫辉光——先命名、再拒绝。
  10. 图片隐含的 brief 必须有真图(餐厅/酒店/时尚/旅行/食品/产品)——零图片是 bug 不是极简。纯文本 + 渐变色块不是 hero。

命令(可选精确入口;自然语言也会触发)

首词匹配命令则按该命令流程走;不匹配但意图清晰也路由过去;无参数则推荐 2-3 个最高价值命令让用户确认,别自动跑。

命令类别做什么主要 reference
craft [brief]构建完整流程:Design Read → dial → substrate → 引擎 → 组装(默认)全部
audit [目标]评估只读诊断:跑反廉价 + spectacle-shown + pre-flight,输出问题列表,不改代码anti-cheap.md preflight.md
redesign [目标]迭代升级已有页面,audit-first,绝不为单个抱怨推倒重来redesign.md
bolder [目标]精修SPECTACLE +2,引擎升一档(Canvas→Three.js)motion-engines.md
quieter [目标]精修SPECTACLE −2,降到 GSAP/CSS-only,压掉过载motion-engines.md
soul [目标]精修页面"感觉很通用"时重挑 personastyle-personas.md
animate [目标]增强只加/换 hero 引擎,只动动效,不碰配色/布局/文案motion-engines.md
densify [目标]增强调 DENSITY ±,增删内容product-ui.md

发布前

说"完成"之前,必须跑 anti-cheap.md 的反廉价扫描 + preflight.md 的完整检查清单。任一硬规则不过 = 发的是破损的活,先修再交付。

© iZiTTMarvin, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 10 other files (references) in .nova/skills/nova-frontend of iZiTTMarvin/Nova-agent.

  • SKILL.md
  • references/anti-cheap.md
  • references/commerce-ui.md
  • references/design-dna.md
  • references/design-systems.md
  • references/motion-engines.md
  • references/patterns.md
  • references/preflight.md
  • references/product-ui.md
  • references/redesign.md
  • references/style-personas.md

Open the folder on GitHubat commit 216466e

Compare with similar skills

Nova Frontend next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Nova Frontend compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Nova Frontend this skilliZiTTMarvin/Nova-agent178—~1.5kAutomated safety check: PassMIT
Design Taste Frontendcaorushizi/oss-client11322 repos~22kAutomated safety check: PassNone
Auteuragiwhitelist/auteur1k—~5kAutomated safety check: PassMIT
Taste Skillyc-software/qm15k—~1.2kAutomated safety check: PassMIT
Design Anti SlopKartikLabhshetwar/better-shot2.4k—~2.5kAutomated safety check: PassCustom licence
Human Gatealirezarezvani/claude-skills28k—~1.5kAutomated safety check: PassMIT

Similar skills

  • Design Taste Frontend

    caorushizi/oss-client

    Anti-slop frontend skill for landing pages, portfolios, and redesigns.

    113 GitHub starsUsed in 22 repos~22k tokens
    Frontend & DesignAuto-check passed
  • Auteur

    agiwhitelist/auteur

    Design and build complete web experiences from scratch — award-level product and marketing pages, cinematic scroll-directed sites where the page is directed like a film, and multi-screen products…

    1k GitHub stars~5k tokensUpdated 2 mo ago
    Frontend & DesignAuto-check passed
  • Taste Skill

    yc-software/qm

    Design taste and process for anything a person will look at in a browser — landing page, dashboard, prototype, deck.

    15k GitHub stars~1.2k tokensUpdated today
    Frontend & DesignAuto-check passed
  • Design Anti Slop

    KartikLabhshetwar/better-shot

    Detect and fix AI design slop — the convergent look that shows up across AI-generated landing pages and dashboards (purple/indigo gradients, rounded-2xl everywhere, three-box feature grids, bento…

    2.4k GitHub stars~2.5k tokensUpdated yesterday
    Frontend & DesignAuto-check passed
  • Human Gate

    alirezarezvani/claude-skills

    Runs the human-verification lane of an agent loop, and proves review happened before work is called done.

    28k GitHub stars~1.5k tokensUpdated 1 mo ago
    Frontend & DesignAuto-check passed
  • UI Craft

    educlopez/ui-craft

    A skill your agent uses for UI design and implementation work to avoid generic AI-looking interfaces.

    375 GitHub stars~9.5k tokensUpdated 1 mo ago
    Frontend & DesignAuto-check passed

More from iZiTTMarvin/Nova-agent

  • Skill Add

    iZiTTMarvin/Nova-agent

    从 URL 或 zip 安装第三方技能到 ~/.nova/skills 或项目 .nova/skills. An agent skill from iZiTTMarvin/Nova-agent.

    178 GitHub stars~277 tokensUpdated yesterday
    Auto-check passed
  • Skill Creator

    iZiTTMarvin/Nova-agent

    交互式创建或优化 Nova Agent 技能(SKILL.md),支持 quick 模式与模板. An agent skill from iZiTTMarvin/Nova-agent.

    178 GitHub stars~355 tokensUpdated yesterday
    Auto-check passed
  • Code Review

    iZiTTMarvin/Nova-agent

    收集工作区未提交改动或最近一次提交的 git 变更面,做只读代码审查并给出带文件行号的问题清单。适用:审查改动、检查刚写的代码质量。不修改代码。

    178 GitHub stars~475 tokensUpdated yesterday
    Auto-check passed
  • Deep Research

    iZiTTMarvin/Nova-agent

    把研究问题拆成子问题后并行只读检索,综合成带来源引用的结论并复核证据。适用:调研主题、技术选型对比、查清事实并给出可追溯结论。不修改代码。

    178 GitHub stars~394 tokensUpdated yesterday
    Auto-check passed
  • New

    iZiTTMarvin/Nova-agent

    空白技能脚手架模板,复制后修改 name/description/body 即可使用. An agent skill from iZiTTMarvin/Nova-agent.

    178 GitHub stars~119 tokensUpdated yesterday
    Auto-check passed
  • Onboard

    iZiTTMarvin/Nova-agent

    首次启动向导:了解工作区、配置规则与技能,快速上手 Nova Agent

    178 GitHub stars~195 tokensUpdated yesterday
    Auto-check passed

Questions about Nova Frontend

What does Nova Frontend do?

Anti-slop, high-craft frontend design: landing pages, brand sites, portfolios, dashboards, product/admin UI, commerce (PDP/PLP/checkout). Nova Frontend is an agent skill from iZiTTMarvin/Nova-agent. Anti-slop, high-craft frontend design: landing pages, brand sites, portfolios, dashboards, product/admin UI, commerce (PDP/PLP/checkout).

When should I use Nova Frontend?

Nova Frontend fits situations like: premium / landing page / dashboard / redesign / give it a soul / /nova-frontend; tasks that involve Landing pages; tasks that involve Humanizing AI text.

How do I install Nova Frontend in Claude Code?

Run `npx skills add iZiTTMarvin/Nova-agent --skill nova-frontend -a claude-code`. Or copy the skill folder (.nova/skills/nova-frontend in iZiTTMarvin/Nova-agent) into .claude/skills/nova-frontend in your project. Claude Code loads it when a task matches its description.

How do I install Nova Frontend in Codex?

Run `npx skills add iZiTTMarvin/Nova-agent --skill nova-frontend -a codex`. Or copy the skill folder (.nova/skills/nova-frontend in iZiTTMarvin/Nova-agent) into .agents/skills/nova-frontend in your project. Codex loads it when a task matches its description.

Can I use Nova Frontend in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add iZiTTMarvin/Nova-agent --skill nova-frontend -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/nova-frontend, .gemini/skills/nova-frontend, .github/skills/nova-frontend and .opencode/skills/nova-frontend in your project.

What does Nova Frontend need to run?

SKILL.md names no scripts, command-line tools or credentials: Nova Frontend is instructions for the agent only.

Does Nova Frontend access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Nova Frontend safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Nova Frontend use?

Nova Frontend is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Nova Frontend use?

About 1.5k tokens (SKILL.md is roughly 6.1k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 24k tokens, read only when the agent opens those files.

What are the alternatives to Nova Frontend?

Skills that share tags, products or a category with Nova Frontend: Design Taste Frontend (caorushizi/oss-client, 113 stars), Auteur (agiwhitelist/auteur, 1k stars), Taste Skill (yc-software/qm, 15k stars) and Design Anti Slop (KartikLabhshetwar/better-shot, 2.4k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Nova Frontend?

iZiTTMarvin (a GitHub user) maintains it in iZiTTMarvin/Nova-agent, which has 178 GitHub stars. The repository holds 7 skills in this directory. The repository was last updated on October 9, 2026.

Source: iZiTTMarvin/Nova-agent on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.