Agent skill

Zui Review

by easysoft in easysoft/zui

只读评审 ZUI 主仓库的指定变更;默认评审本地未提交改动,没有改动时评审未推送提交。适用于 diff 评审,不用于整库优化审计。

MITAuto-check passedFrontend & Design

Install Zui Review

skills CLI
$ npx skills add easysoft/zui --skill zui-review -a claude-code

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

GitHub CLI
$ gh skill install easysoft/zui zui-review --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/easysoft/zui.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.agents/skills/zui-review .claude/skills/zui-review && 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
zui-review
GitHub stars
2.8k
Token cost
~1.3k tokens
SKILL.md length
321 words
Files
2
Skills in repo
24
Repo updated
First seen
Licence
MIT

At a glance

只读评审 ZUI 主仓库的指定变更;默认评审本地未提交改动,没有改动时评审未推送提交。适用于 diff 评审,不用于整库优化审计。

  • Works in 3 steps: 用户指定范围:严格使用用户点名的暂存层级、提交、commit… → 本地未提交改动:用户未指定变更来源时,先运行 git… → 本地未推送提交:只有当前范围没有未提交改动时,优先解析当前分支的 @{push}…
  • Frontend & Design work in your project
  • SKILL.md covers 操作边界, 确定评审范围, 评审方法 and 严重度, plus 1 more section
  • Calls git and pnpm

What it does

Zui Review is an agent skill from easysoft/zui. 只读评审 ZUI 主仓库的指定变更;默认评审本地未提交改动,没有改动时评审未推送提交。适用于 diff 评审,不用于整库优化审计。

Its SKILL.md is about 1.3k tokens, which your agent loads only when the skill is triggered. The skill folder holds 2 other files (for example `agents/openai.yaml`).

It sits in Frontend & Design. It works with Git. The repository describes itself as: ZUI is an HTML5 front UI framework. The licence is MIT.

When your agent uses it

  • Frontend & Design work in your project

Example prompts

  • “/zui-review”

Workflow steps

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

  1. 用户指定范围:严格使用用户点名的暂存层级、提交、commit range、分支、diff 或 PR 作为变更来源,解析并验证引用,但不混入其他来源;保留用户明确给出的 .. / ... 语义。单个普通提交按 parent-to-commit 评审;单个 merge…
  2. 本地未提交改动:用户未指定变更来源时,先运行 git --no-optional-locks status --short --branch,再用 git --no-optional-locks status --porcelain=v1…
  3. 本地未推送提交:只有当前范围没有未提交改动时,优先解析当前分支的 @{push} 作为基线。用 git rev-list --left-right --count ...HEAD 记录 behind/ahead,用 git log --reverse ..HEAD…

What it can do on your machine

Read from SKILL.md and the folder at commit 59b9aeb. 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

    Shell commands in SKILL.md call:

    • git
    • pnpm

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

  • Network

    No URLs in SKILL.md. Its commands use git and pnpm, which can reach the network depending on how they are called.

    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

Zui Review loads about 1.3k tokens when it runs. Until then it costs about 19 tokens; SKILL.md has 321 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~19
When it runs · the whole SKILL.md, loaded when a task matches
~1.3k

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 easysoft/zui at commit 59b9aeb, republished under its MIT licence (© easysoft). 321 words, ~1,306 tokens.

Download SKILL.mdSave it as .claude/skills/zui-review/SKILL.md (or your agent's skills folder). This skill also uses 1 other file; get the full folder from GitHub.
name
zui-review
description
只读评审 ZUI 主仓库的指定变更;默认评审本地未提交改动,没有改动时评审未推送提交。适用于 diff 评审,不用于整库优化审计。

ZUI 代码评审

操作边界

  • 评审保持只读;“review”“检查代码”或“评审代码”不表示授权修复、格式化、暂存、提交、推送、切换分支或更新远端引用。
  • 用户另行明确要求修复时,先完成评审,再按变更所属的 ZUI 实施技能直接处理请求范围内的修复;不要让修复诉求改变本次评审范围。
  • 保留用户已有的暂存、未暂存和未跟踪内容。不要运行 --fix、会刷新生成目录或缓存的验证命令;用户明确要求相应验证或能够可靠使用仓库外临时输出时除外。
  • 读取目标适用的 AGENTS.md;已有且未变化的上下文直接复用。涉及 lib/* 时,按 ZUI 公共规范 和 协作工作流 读取判断本次变更所需的规范与源码。

确定评审范围

按以下优先级选择唯一范围,并在输出中说明最终使用的范围和基线:

  1. 用户指定范围:严格使用用户点名的暂存层级、提交、commit range、分支、diff 或 PR 作为变更来源,解析并验证引用,但不混入其他来源;保留用户明确给出的 .. / ... 语义。单个普通提交按 parent-to-commit 评审;单个 merge commit 先检查其父提交,若用户需要的 parent 或 combined 语义仍不明确则询问,不擅自固定为 first-parent。用户只点名文件、目录或库时,将其作为 path filter,并在该 filter 内按下面规则确定变更来源;把 pathspec 追加到每个适用命令的 -- 后。范围仍有多种实质解释且无法从上下文消除时,先提出最少必要问题。
  2. 本地未提交改动:用户未指定变更来源时,先运行 git --no-optional-locks status --short --branch,再用 git --no-optional-locks status --porcelain=v1 --untracked-files=all -- <pathspec> 判断当前范围;没有 path filter 时省略末尾的 -- <pathspec>。只要该 porcelain 存在 staged、unstaged、untracked 或冲突条目,就评审它们的并集:
    • 分别读取 git diff --cached --no-ext-diff --find-renames 与 git diff --no-ext-diff --find-renames,保留暂存和未暂存边界;
    • HEAD 可解析时,使用 git diff HEAD 理解 tracked 文件的最终组合状态;unborn branch 跳过该命令,以 cached、unstaged 和逐文件读取重建状态;
    • 使用 git ls-files --others --exclude-standard -z 枚举并逐个读取未跟踪文件,因为普通 diff 不包含它们;
    • 对部分暂存、重命名、删除、冲突和二进制文件单独确认真实状态,不因某一种 diff 为空而漏审。
  3. 本地未推送提交:只有当前范围没有未提交改动时,优先解析当前分支的 @{push} 作为基线。用 git rev-list --left-right --count <base>...HEAD 记录 behind/ahead,用 git log --reverse <base>..HEAD -- <pathspec> 枚举当前 filter 内仅在本地可达且有相关改动的提交,并以 git diff --no-ext-diff --find-renames <base>...HEAD -- <pathspec> 评审相对 merge-base 的聚合结果;没有 path filter 时省略末尾的 -- <pathspec>。需要理解被聚合 diff 隐藏的重排、回退或提交意图时,再读取单个提交。

如果 @{push} 无法解析但 @{upstream} 能可靠解析,可以退回 upstream,并明确说明它不一定等于实际推送目标;两者都不可用时再请用户给出基线。遇到 detached HEAD、unborn branch、历史不完整、没有共同祖先或引用无法解析时,不要猜测范围。ahead 为零,或 path filter 下的聚合 diff 为空,表示没有默认评审内容,直接报告这一事实;不要退而评审最近一次提交。不要擅自改用 origin/dev、默认分支或整个历史。分支已分叉时明确报告 ahead/behind 状态,但仍只评审基线两点范围中仅在本地可达的提交。push/upstream 远端跟踪引用只是本地缓存;默认不运行 git fetch,需要确认远端最新状态时必须由用户明确要求。

ignored 文件不属于默认未提交范围,用户点名时才纳入。显式 path filter 可能被普通 status 和 --exclude-standard 隐藏;先用 git check-ignore -- <pathspec> 识别,再以 git ls-files --others --ignored --exclude-standard -z -- <pathspec> 或直接逐文件读取,只纳入用户点名的 ignored 目标。子模块默认只评审父仓库记录的 gitlink 变化,不自动递归 dirty 子模块或嵌套仓库。所有 pathspec 放在 -- 后;不要通过 eval 执行用户提供的 revision 或路径。

评审方法

  1. 读取完整 diff、变更文件及必要上下文,不只看改动行。追踪相关入口、公共类型、调用方、消费者、测试/调试页、配置和生成源,确认变更实际进入的执行路径与公开契约。
  2. 重点检查本次变更是否引入或暴露以下问题:
    • 逻辑错误、错误恢复缺口、空值与边界条件、异步竞态、重复执行和状态不同步;
    • 公共 API、类型、事件、DOM/CSS、序列化、包元数据或向后兼容性回归;
    • 本次新增、替换或升级的第三方依赖及资源,按版权与文档规则核对版权文件、适用版本与公共依赖说明;保留开发依赖部分,不因外置加载或声明在 devDependencies 中而漏检运行时资源;
    • listener、timer、observer、portal、实例、缓存及其他资源的初始化、更新与销毁不对称;
    • 安全、数据损坏、无障碍、国际化及有可复现场景的性能退化;
    • Preact 组件违反状态与副作用规范:使用 hooks(含 signals hooks),或未在卸载/销毁时清理 effect;
    • Web Component 变更按 专用规范 核对具体组件定义是否实际位于 lib/<lib-name>/src/web-component/、是否经目录 index.ts 和 src/main.ts 导出,以及注册时机、渲染目标推导、配置缓存、属性反射及断开清理;
    • ZUI 的 Preact/vanilla 双形态、组件注册、@zui/* 跨库导入、contributes、主题和 HMR 约定;
    • 导入、公开入口、类型或文档接入变化按源码与文档接入规范核实:是否把工作区入口当作 npm 入口、虚构子路径或组件形态、混用类型与运行时成员、依赖仅在文档站存在的环境,或在接入修复中夹带版本变化;
    • 组件 CSS 与样式入口变更按 布局与样式规范 核实 @apply 优先、style/index.ts 导入完整性、层叠顺序及消费入口;遵循原生 CSS 的适用情形,不把等价写法差异当作运行时缺陷;
    • 布局或示例类名变更按 CSS utilities 核实流程 检查:是否照搬未公开的 Tailwind 类、变体或任意值,误用同名类的默认值或语义,或依赖仅在开发/文档环境生成的样式;结合组件自定义样式和实际消费入口确认影响,不能仅因 utilities 中没有该类就报告缺陷;
    • 新建可视组件库、改变主要外观/用途或修改封面时,按组件预览规范检查标准预览是否缺失、过时或违反静态/主题契约;不把与当前 diff 无关的历史缺图列为 finding;
    • 生成文件与 source-of-truth 不一致。遇到生成产物时找到生成器或映射并审查源头,不只评审生成结果。
  3. 搜索相关符号、相似成熟实现和调用点来验证判断。发现看似异常的代码时,先确认是否为现有约定、兼容处理或基线问题。
  4. 对 tracked Git diff 运行对应范围的 git diff --check。它不覆盖未跟踪文件;对未跟踪文本文件使用等价 whitespace 检查,或以 git diff --no-index --check -- /dev/null <file> 检查诊断内容,并注意 no-index 因“存在内容差异”返回非零不等于 whitespace 失败。按风险选择不会修改工作区的目标 lint、类型或现有测试命令;先从当前 package.json 确认可用脚本,不臆造 pnpm test。lint 输出先作为验证结果,只有它证明本次变更引入真实缺陷或确定阻断既有门禁时才升级为 finding。把静态检查、构建、浏览器/runtime 验证分开报告。无法安全运行的验证说明原因和剩余风险,不把“未运行”写成“通过”。
  5. 只报告由当前范围引入、可操作且有充分证据的问题。不要把以下内容列为 finding:纯风格偏好、没有具体后果的防御性建议、范围外历史缺陷、与改动无关的基线失败,或仅凭猜测成立的风险。测试缺口只有在明确导致关键行为无法验证或违反现有契约时才作为 finding;否则放入残余风险。

严重度

  • P0:阻断发布或核心使用,或造成广泛安全问题、数据损坏、构建完全失败。
  • P1:常见路径上的明确运行时故障、公共 API/兼容性回归或严重生命周期问题。
  • P2:有现实触发条件的重要边界错误、资源泄漏、无障碍问题或显著行为偏差。
  • P3:影响有限但真实、可操作的缺陷;不要用 P3 容纳风格和可选改进。

严重度按影响和触发概率判断,不按改动大小、文件数量或修复难度判断。

输出

  1. 先列 findings,按 P0 到 P3 排序。每项只描述一个根因,使用简短标题,并包含最窄的文件/行号、触发场景、实际影响和判断依据;行号尽量落在当前评审终点的改动行上。纯删除导致的问题锚定最近的存续改动行并说明删除关系。宿主支持行内评论时使用行内评论,否则使用可点击路径与行号。
  2. 不确定但值得关注的事项单列为风险或问题,不与已确认 findings 混排。
  3. 没有 finding 时明确写“未发现可操作问题”,不要为了显得完整而制造建议。
  4. 最后简要列出已评审范围、运行过的验证及结果、未覆盖的 runtime/浏览器场景和其他残余风险。不要用 lint、构建或测试通过代替代码评审结论。

© easysoft, 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 1 other file in .agents/skills/zui-review of easysoft/zui.

  • SKILL.md
  • agents/openai.yaml

Open the folder on GitHubat commit 59b9aeb

Compare with similar skills

Zui Review 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.

Zui Review compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Zui Review this skilleasysoft/zui2.8k—~1.3kAutomated safety check: PassMIT
Port Componenttheexperiencecompany/gaia-ui220—~2.9kAutomated safety check: PassMIT
Nutui Build Local Verifyjdf2e/nutui-react1.2k—~472Automated safety check: PassNone
Compare Array Bundle SizePostHog/posthog-js628—~599Automated safety check: PassCustom licence
UX Create Manifesth0x91b/dev-3.0307—~2.7kAutomated safety check: PassApache-2.0
Verify Atlaspoteto/verification-skill-example134—~1.4kAutomated safety check: PassNone

Similar skills

  • Port Component

    theexperiencecompany/gaia-ui

    Ports a UI component from the gaia mono repo (theexperiencecompany/gaia) into the gaia-ui registry.

    220 GitHub stars~2.9k tokensUpdated 1 mo ago
    Frontend & DesignAuto-check passed
  • Nutui Build Local Verify

    jdf2e/nutui-react

    NutUI 比例缩放本地验证——写回 src/packages 下同路径组件 SCSS(跳过 src/packages//demo.scss 与 demos);--mirror 写 scale-verify/;不写 build。

    1.2k GitHub stars~472 tokensUpdated today
    Frontend & DesignAuto-check passed
  • Compare Array Bundle Size

    PostHog/posthog-js

    Official

    Quickly compare the posthog-js array.js bundle size in the current working tree against a git baseline using the repository's esbuild proxy.

    628 GitHub stars~599 tokensUpdated today
    Frontend & DesignAuto-check passed
  • UX Create Manifest

    h0x91b/dev-3.0

    Create the initial Product UX Bible for an existing web or full-screen web app by deeply auditing the repository, using sub-agents when available, and generating docs/ux manifests, schemas, budgets…

    307 GitHub stars~2.7k tokensUpdated yesterday
    Frontend & DesignAuto-check passed
  • Verify Atlas

    poteto/verification-skill-example

    Drive the Atlas workspace UI in a running Harbor Labs desktop build via CDP.

    134 GitHub stars~1.4k tokensUpdated 2 mo ago
    Frontend & DesignAuto-check passed
  • Worklog Design

    regisx001/Worklog

    Design and UI skill for the Worklog desktop project manager.

    261 GitHub stars~3.3k tokensUpdated 1 mo ago
    Frontend & DesignAuto-check passed

More from easysoft/zui

All 24 skills in this repo
  • Zui Build

    easysoft/zui

    使用内置 ZUI 资源从零创建可运行的独立网页或小型静态站点。向已有应用集成 ZUI 时使用 zui 技能. An agent skill from easysoft/zui.

    2.8k GitHub stars~1.4k tokensUpdated today
    Auto-check passed
  • Zui

    easysoft/zui

    在已有应用中集成、使用或排查 ZUI 3,直接创建和注册应用自定义组件,或使用已提供的 Web Components;适用于消费方项目,不用于开发 ZUI 源码仓库。

    2.8k GitHub stars~2.2k tokensUpdated today
    Auto-check passed
  • Zui Standards

    easysoft/zui

    查询 ZUI 主仓库的开发规范,为库开发及本地 exts 接入任务提供相关约束和仓库证据;规范查询保持只读. An agent skill from easysoft/zui.

    2.8k GitHub stars~452 tokensUpdated today
    Auto-check passed
  • Zuix Standards

    easysoft/zui

    查询独立 ZUI 扩展项目的开发规范,解析目标库、扩展工作区、Git 所有权及所需宿主上下文;始终只读. An agent skill from easysoft/zui.

    2.8k GitHub stars~403 tokensUpdated today
    Auto-check passed
  • Zui Wrap Lib

    easysoft/zui

    将已经可用的 UMD/IIFE 资源封装为 ZUI 主仓库的 lib/ 库,通过 LibLoader 按需加载。不转换资源格式,需求明确时形成集成计划后直接实施。

    2.8k GitHub stars~1k tokensUpdated today
    Auto-check passed
  • Zui Commit

    easysoft/zui

    审查并提交 ZUI 主仓库的改动;用户直接调用本技能即进入提交模式,必要检查通过后直接提交。只要求分析、评审或生成提交信息时保持只读。

    2.8k GitHub stars~1.5k tokensUpdated today
    Auto-check: notes

Works with

Questions about Zui Review

What does Zui Review do?

只读评审 ZUI 主仓库的指定变更;默认评审本地未提交改动,没有改动时评审未推送提交。适用于 diff 评审,不用于整库优化审计。. Zui Review is an agent skill from easysoft/zui.

When should I use Zui Review?

Zui Review fits situations like: frontend & Design work in your project.

How do I install Zui Review in Claude Code?

Run `npx skills add easysoft/zui --skill zui-review -a claude-code`. Or copy the skill folder (.agents/skills/zui-review in easysoft/zui) into .claude/skills/zui-review in your project. Claude Code loads it when a task matches its description.

How do I install Zui Review in Codex?

Run `npx skills add easysoft/zui --skill zui-review -a codex`. Or copy the skill folder (.agents/skills/zui-review in easysoft/zui) into .agents/skills/zui-review in your project. Codex loads it when a task matches its description.

Can I use Zui Review 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 easysoft/zui --skill zui-review -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/zui-review, .gemini/skills/zui-review, .github/skills/zui-review and .opencode/skills/zui-review in your project.

What does Zui Review need to run?

Going by SKILL.md and its folder, Zui Review needs the command-line tools its instructions call (git and pnpm).

Does Zui Review access the network?

SKILL.md contains no URLs. Its commands use git, which can reach the network depending on how they are called. This is read from the text; nothing was executed.

Is Zui Review 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 Zui Review use?

Zui Review 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 Zui Review use?

About 1.3k tokens (SKILL.md is roughly 5.2k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Zui Review?

Skills that share tags, products or a category with Zui Review: Port Component (theexperiencecompany/gaia-ui, 220 stars), Nutui Build Local Verify (jdf2e/nutui-react, 1.2k stars), Compare Array Bundle Size (PostHog/posthog-js, 628 stars) and UX Create Manifest (h0x91b/dev-3.0, 307 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Zui Review?

easysoft (a GitHub organization) maintains it in easysoft/zui, which has 2,766 GitHub stars. The repository holds 24 skills in this directory. The repository was last updated on October 9, 2026.

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