← Back to Skills Marketplace
110
Downloads
0
Stars
0
Active Installs
2
Versions
Install in OpenClaw
/install huo15-openclaw-design-critique
Description
5 维度设计评审（美学 / 可用性 / 品牌一致 / 内容 / 技术实现）+ Keep/Fix/Quick Wins 三分类 + 雷达图结构化输出。用于对现成网页、HTML、截图、Figma 链接做专业打分与改进建议。触发词：设计评审、UI 审查、给这个页面打分、review 这个设计、这个界面怎么样、优化建议。
README (SKILL.md)
火一五设计评审技能 v1.0
结构化 5 维设计评审 — 青岛火一五信息科技有限公司
一、触发场景
当用户对一个已存在的界面/页面/截图，要求专业评价或改进建议：
- "给这个页面打分"
- "review 一下这个 UI"
- "这个设计哪里有问题"
- "这个落地页怎么优化"
- 附带一张截图或一个 URL 要求评估
产出：5 维打分（雷达图文本化）+ Keep/Fix/Quick Wins 三分类清单 + 优先级排序。
二、五维评分体系（每项 1-5 分）
| 维度 | 关注点 | 1 分 | 5 分 |
|---|---|---|---|
| 美学 Aesthetic | 字体 / 颜色 / 动效 / 空间 / 氛围 | 通用 AI slop | 意图明确的流派 |
| 可用性 Usability | 信息层级、可点击性、反馈、a11y | 用户需要猜 | 一眼理解下一步 |
| 品牌一致 Brand | 与品牌调性 / 产品定位一致度 | 和任何其他站能混用 | 一眼辨识品牌 |
| 内容 Content | 文案质量、信息密度、有无废话 | 占位文案 / 废话堆砌 | 每个字都在做事 |
| 实现 Implementation | 性能、响应式、代码质量、可维护 | 打开慢 / 布局崩 | 丝滑 / 鲁棒 |
总分规则：
- 5×5 = 25 分制
- 短板决定印象：任何一项 ≤ 2 分，总分不得超过 15（木桶短板）
- "AI slop 雷区"：命中
huo15-openclaw-frontend-design §四硬红线任一，美学直接 ≤ 2
三、输出结构（每次评审必按此格式）
### 总评
[一句话总结：这个设计是什么流派？最大亮点？最大问题？]
### 五维打分
| 维度 | 分数 | 备注 |
|------|------|------|
| 美学 | X/5 | ... |
| 可用性 | X/5 | ... |
| 品牌一致 | X/5 | ... |
| 内容 | X/5 | ... |
| 实现 | X/5 | ... |
**总分：XX/25**
### 雷达图（ASCII）
美学 5
●
内容 ● ─┼─ ● 可用性
●│●
实现 ● 品牌
(用 ● 的位置表示分数，1-5 对应距离中心)
### Keep（保持）
- ...（这几点做得好，不要动）
### Fix（必改）
- P0 ...（上线前必须修）
- P1 ...（下个迭代）
### Quick Wins（1 小时内能做）
- ...（性价比最高的小改动）
### 改进后预期
[预计改完能从 XX/25 到 YY/25]
四、评审工作流
阶段 1 · 获取素材
- 如果给的是 URL → 要求用户提供截图（调用
huo15-openclaw-frontend-design §六·阶段 4的 Playwright 截图 CLI） - 如果给的是 HTML 代码 → 自己读 + 要求用户在本地浏览器打开后截图
- 如果只给了描述 → 拒绝评审，要求至少提供一张截图
阶段 2 · 硬红线体检（30 秒）
对照 huo15-openclaw-frontend-design §四 的 8 条硬红线逐一过一遍。命中直接列入 Fix · P0。
阶段 3 · 五维打分
依次给 5 个维度打分，每个维度至少一条具体理由（不能只给分数）。
阶段 4 · 三分类建议
- Keep：避免改掉的闪光点（保护用户已有投入）
- Fix：必改项，按 P0/P1 排序
- Quick Wins：投入 ≤ 1 小时、收益明显的
阶段 5 · 出报告
按 §三的格式出一份完整报告。
五、常见评审反模式（Claude 自己要避免）
- ❌ 只夸不改 —— 这不是评审是吹捧
- ❌ 全盘否定 —— 用户的投入也有价值，要找 Keep
- ❌ 空话大话 —— "这里可以更现代化"，没用。要说"字体改成 Playfair Display，字号从 48px 改到 64px"
- ❌ 建议互相矛盾 —— 先想清楚整体方向再逐条提
- ❌ 忽略实现成本 —— 要分 P0/P1/Quick Win
六、与其他 huo15 技能的分工
| 场景 | 归属 |
|---|---|
| 评审 Web UI / HTML / 截图 | 本技能 |
| 评审 PPT 演示稿 | huo15-openclaw-ppt 的 review 模式（如有） |
| 评审 Word 文档结构 | huo15-openclaw-office-doc |
| 生成设计后自检 | huo15-openclaw-frontend-design §六·阶段 5 调用本技能 |
七、触发词
- 设计评审 / UI 审查 / UX 评审
- 给这个页面打分 / 给这个设计打分
- review 这个设计 / review UI
- 这个界面怎么样 / 这个页面有什么问题
- 优化建议 / 改进建议（针对已有页面）
八、版本历史
- v1.0.0（当前 · 2026-04-23）：初始版本。5 维评分体系 + Keep/Fix/Quick Wins 三分类 + ASCII 雷达图 + 硬红线体检快速筛 + 5 类反模式警示。
技术支持： 青岛火一五信息科技有限公司
Usage Guidance
This skill otherwise looks like a reasonable design-review prompt set, but it references running a Playwright screenshot CLI and other 'huo15' skill sections while declaring no binaries or installs. Before installing or enabling: 1) Confirm whether your agent environment provides Playwright or the referenced 'huo15-openclaw-frontend-design' capability. If not, insist the skill author declare required binaries or provide an alternative (accept user-uploaded screenshots). 2) If you allow the agent to run Playwright or any CLI, ensure it will only access pages you explicitly permit (avoid letting it browse intranet or authenticated pages without review). 3) Prefer providing screenshots or sanitized HTML yourself rather than granting the agent access to local files or automated browsing. 4) Ask the publisher for a clear dependency list and explicit privacy guarantees (what data is sent where) — absence of declared dependencies here is the main reason this is flagged as suspicious.
Capability Assessment
Purpose & Capability
The skill's name, description, and detailed SKILL.md all align with a design critique capability (5-dimension scoring, Keep/Fix/Quick Wins, ASCII radar). However the workflow explicitly references calling a Playwright screenshot CLI and cross-calls to other 'huo15-openclaw-frontend-design' skill sections; the package metadata declares no required binaries or install steps. Either the agent will expect an external Playwright binary or another skill to exist — the absent dependency is disproportionate to the claimed 'instruction-only' nature and should be declared.
Instruction Scope
The runtime instructions are prescriptive and generally within the domain of a design review. Concerns: (1) Stage 1 instructs the agent to call an external Playwright CLI from another huo15 skill for screenshots — this is a runtime external command not declared in requirements; (2) instructions say 'if given HTML code → read it' and request the user open it locally and screenshot, which pushes data collection onto the user but also implies the agent may attempt to access or render HTML; (3) the skill cross-references internal '硬红线' rules and other skills, which assumes those other skills or documents exist. There is no instruction to access secrets or system config, and the skill refuses purely textual descriptions, which is a reasonable scope restriction.
Install Mechanism
This is an instruction-only skill with no install spec and no code files, which is the lowest-risk install model. There are no downloads, packages, or extracted archives. The only concern is that the instructions expect external tooling (Playwright) which is not declared as required — but there is no install mechanism in this skill itself.
Credentials
The skill requests no environment variables, no credentials, and no config paths. That is proportionate to a read-only review task. There is no attempt to access unrelated secrets or system files in the SKILL.md.
Persistence & Privilege
always is false and disable-model-invocation is false (normal). The skill does not request permanent presence or elevated privileges and does not instruct modifications to other skills' configs. No persistence/privilege concerns were found.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install huo15-openclaw-design-critique - After installation, invoke the skill by name or use
/huo15-openclaw-design-critique - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.1
- No changes detected in this version.
- Version number and content remain the same as the previous release.
v1.0.0
首发 v1.0.0：5 维评审 + Keep/Fix/Quick Wins 三分类 + 木桶短板总分 + ASCII 雷达图
Metadata
Frequently Asked Questions
What is Huo15 Openclaw Design Critique?
5 维度设计评审（美学 / 可用性 / 品牌一致 / 内容 / 技术实现）+ Keep/Fix/Quick Wins 三分类 + 雷达图结构化输出。用于对现成网页、HTML、截图、Figma 链接做专业打分与改进建议。触发词：设计评审、UI 审查、给这个页面打分、review 这个设计、这个界面怎么样、优化建议。 It is an AI Agent Skill for Claude Code / OpenClaw, with 110 downloads so far.
How do I install Huo15 Openclaw Design Critique?
Run "/install huo15-openclaw-design-critique" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Huo15 Openclaw Design Critique free?
Yes, Huo15 Openclaw Design Critique is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Huo15 Openclaw Design Critique support?
Huo15 Openclaw Design Critique is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Huo15 Openclaw Design Critique?
It is built and maintained by Job Zhao (@zhaobod1); the current version is v1.0.1.
More Skills