Create a new project-specific skill in .claude/skills/ to codify a recurring pattern.
English\"Comprehensive frontend analysis including DOM, CSS, layout, and structure. 全面析前端含DOM樣式布局結構.
English\"Analyze user experience including layout, interactions, and usability. 析用戶體驗含布局交互可用性.
English\"Analyze page performance including load times and network resources. 析頁性能含載時網絡資源.
EnglishRecord + replay interactive guided DEMO of shipped work — floating self-advancing walkthrough narrating each step, spotlighting live app element, advancing on timer, click, or app-state condition. 在瀏覽器疊層錄製可重播之互動導覽。 Use when: demo a feature, create walkthrough, run walkthrough tool, guided product tour, replay a flow, show what was built, onboarding tour, step-by-step demo overlay.
EnglishRoutes caveman / context-compression intent to right manual skill — sole auto gateway; others manual. caveman 唯一自動網關。 Use when: compress file/PDF/slides/spreadsheet/audio, condense git diff/log or PR URL, OCR/summarize image/screenshot/diagram, compress error logs/stack traces/notes, fetch+compress URL (YouTube/GitHub/arXiv), multi-source research, caveman-mcp setup, unsure which caveman skill.
EnglishWrite value-first commit message (subject + body) for staged/unstaged changes or branch work. Use when: writing commit message, about to stage changes, before commit, drafting commit body, summarizing branch work for commit. Also: writing PR description when user explicitly invokes --pr or says 'for the PR'. Returns title + body; never runs git commit / gh pr create itself.
EnglishOverview of a11y-audit MCP server tools for static accessibility analysis - HTML audits, WCAG scoring, heading outlines, ARIA validation, link text checks, and PDF accessibility inspection without browser.
EnglishArchitecture/security/performance claims need >=2 independent sources or escalate. Load-bearing = wrong cascades downstream. Use when: spec/ADR architectural decisions, security primitive choice, performance characteristic assertion, PR with load-bearing claim, subagent design doc. Skip: typos/comments, single authoritative source documented, exploratory hypothesis.
EnglishNew memory entries need ISO 8601 timestamp + source-event tag (loop-id/task-id/commit-sha) for temporal conflict resolution. Soft guidance — backfill on next touch. Use when: writing new memory, creating per-memory file, reviewing PR adding memory, subagent appending memory. Skip: existing files lacking fields, CLAUDE.md preferences.
EnglishAdversarial cooperation loop for creating complete design library from Figma with CSS, assets, HTML mockups, and documentation.
EnglishValidate this plugin marketplace before committing — checks plugin manifests, hook wiring, skill/command/agent frontmatter, and marketplace.json consistency. Use when adding or editing a plugin, a skill, a command, an agent, or hooks; when a plugin's hooks or skills mysteriously do not load; or before any commit or release to this repo.
English\"Analyze prompts for clarity, effectiveness, and optimization opportunities.
EnglishResolve review feedback by evaluating validity and fixing issues in parallel. Pre-commit-first. Use when: applying review feedback from local reviewer agent output, resolving pre-commit review findings, working through code-review agent suggestions, addressing reviewer-flagged issues before commit. Also: PR comments via gh as secondary path.
EnglishProblem-first discovery: given a goal, find a registered slop-mcp tool that solves it, verify its schema, then run it. 由問題出發,覓已注冊 MCP 工具解之,驗 schema 而後執行。 Use when: is there a tool for X, find a tool to do X, solve a task via a registered MCP server, unsure which MCP capability exists, capability discovery before building custom.
EnglishRoutes any slop-mcp / MCP intent to the one right manual slop skill — the single auto-invoked gateway; all other slop skills are manual. slop-mcp 唯一自動網關,導向正確之手動技藝。 Use when: mention MCP or slop, unsure which slop skill, want a registered MCP capability, register/discover/execute/customize/monitor an MCP server, import configs, MCP memory or scripting, before shelling out to an MCP binary.
EnglishFull adversarial quality loop — implement, self-attack, parallel verification, quality gates, final validation. 全品質循環:實現、自攻、並行驗證、質量關卡、最終確認.
English
AGENT SKILLS ATLASIncludes Agnt Analyze Frontend, Agnt Analyze UX, Agnt Audit Accessibility, and other workflows.
Claude Code
Codex
Cursor
GitHub Copilot
Gemini CLISupported by the Atlas install prompt. Skill-specific requirements vary.
Pack stats
What’s included
Add Skill
Create a new project-specific skill in .claude/skills/ to codify a recurring pattern. 在項目中新建可複用技能。 Use when: add a skill, create a skill, new skill, codify a pattern, make a reusable workflow
Agnt Analyze Frontend
\"Comprehensive frontend analysis including DOM, CSS, layout, and structure. 全面析前端含DOM樣式布局結構. Use when: analyze frontend code, inspect DOM structure, review CSS layout, audit page structure, diagnose rendering issues\
Agnt Analyze UX
\"Analyze user experience including layout, interactions, and usability. 析用戶體驗含布局交互可用性. Use when: review UX, analyze interactions, check usability, audit user flows, evaluate interface design\
Agnt Audit Accessibility
\"Run comprehensive accessibility audit on current page. 全面審現頁無障礙性. Use when: audit accessibility, check WCAG compliance, find a11y issues, test screen reader support, review ARIA labels\
Agnt Audit Performance
\"Analyze page performance including load times and network resources. 析頁性能含載時網絡資源. Use when: check page speed, audit load times, analyze network waterfall, optimize performance, find slow resources\
Agnt Audit SEO
\"Audit page for SEO best practices and issues. 審頁面搜尋優化得失. Use when: check SEO score, fix meta tags, audit page titles, review structured data, improve search ranking\
Agnt Demo Flow
Record + replay interactive guided DEMO of shipped work — floating self-advancing walkthrough narrating each step, spotlighting live app element, advancing on timer, click, or app-state condition. 在瀏覽器疊層錄製可重播之互動導覽。 Use when: demo a feature, create walkthrough, run walkthrough tool, guided product tour, replay a flow, show what was built, onboarding tour, step-by-step demo overlay.
Caveman Recommender
Routes caveman / context-compression intent to right manual skill — sole auto gateway; others manual. caveman 唯一自動網關。 Use when: compress file/PDF/slides/spreadsheet/audio, condense git diff/log or PR URL, OCR/summarize image/screenshot/diagram, compress error logs/stack traces/notes, fetch+compress URL (YouTube/GitHub/arXiv), multi-source research, caveman-mcp setup, unsure which caveman skill.
Commit Description
Write value-first commit message (subject + body) for staged/unstaged changes or branch work. Use when: writing commit message, about to stage changes, before commit, drafting commit body, summarizing branch work for commit. Also: writing PR description when user explicitly invokes --pr or says 'for the PR'. Returns title + body; never runs git commit / gh pr create itself.
Design Utilities Accessibility Audit Overview
Overview of a11y-audit MCP server tools for static accessibility analysis - HTML audits, WCAG scoring, heading outlines, ARIA validation, link text checks, and PDF accessibility inspection without browser.
Dev Standards Load Bearing Sources
Architecture/security/performance claims need >=2 independent sources or escalate. Load-bearing = wrong cascades downstream. Use when: spec/ADR architectural decisions, security primitive choice, performance characteristic assertion, PR with load-bearing claim, subagent design doc. Skip: typos/comments, single authoritative source documented, exploratory hypothesis.
Dev Standards Memory Needs Source
New memory entries need ISO 8601 timestamp + source-event tag (loop-id/task-id/commit-sha) for temporal conflict resolution. Soft guidance — backfill on next touch. Use when: writing new memory, creating per-memory file, reviewing PR adding memory, subagent appending memory. Skip: existing files lacking fields, CLAUDE.md preferences.
Figma Query Adversarial Design Library
Adversarial cooperation loop for creating complete design library from Figma with CSS, assets, HTML mockups, and documentation. 對抗式協作循環,從 Figma 生成完整設計庫:CSS、資產、HTML、文檔。 Use when: extracting complete design library, running adversarial quality verification loop, creating production-ready component library from Figma, full design-to-code extraction
Marketplace Validation
Validate this plugin marketplace before committing — checks plugin manifests, hook wiring, skill/command/agent frontmatter, and marketplace.json consistency. Use when adding or editing a plugin, a skill, a command, an agent, or hooks; when a plugin's hooks or skills mysteriously do not load; or before any commit or release to this repo.
Prompt Engineer Analyze Prompt
\"Analyze prompts for clarity, effectiveness, and optimization opportunities. 分析提示之清晰度、有效性及優化機會。 Use when: reviewing existing prompts, identifying issues before deployment, generating improvement recommendations.\
Resolve Review Feedback
Resolve review feedback by evaluating validity and fixing issues in parallel. Pre-commit-first. Use when: applying review feedback from local reviewer agent output, resolving pre-commit review findings, working through code-review agent suggestions, addressing reviewer-flagged issues before commit. Also: PR comments via gh as secondary path. 解決審查回饋:本地審查代理輸出優先,PR次之。
Slop MCP Find Tools
Problem-first discovery: given a goal, find a registered slop-mcp tool that solves it, verify its schema, then run it. 由問題出發,覓已注冊 MCP 工具解之,驗 schema 而後執行。 Use when: is there a tool for X, find a tool to do X, solve a task via a registered MCP server, unsure which MCP capability exists, capability discovery before building custom.
Slop MCP Recommender
Routes any slop-mcp / MCP intent to the one right manual slop skill — the single auto-invoked gateway; all other slop skills are manual. slop-mcp 唯一自動網關,導向正確之手動技藝。 Use when: mention MCP or slop, unsure which slop skill, want a registered MCP capability, register/discover/execute/customize/monitor an MCP server, import configs, MCP memory or scripting, before shelling out to an MCP binary.
Workflow Adversarial Quality
Full adversarial quality loop — implement, self-attack, parallel verification, quality gates, final validation. 全品質循環:實現、自攻、並行驗證、質量關卡、最終確認. Use when: implement task with quality, run adversarial review, verify code quality, run quality gates, complete workflow task
Quick start
- Choose your skills. Select individual skills above, or choose the entire pack.
- Copy the install prompt. Use Copy install prompt above the list.
- Send it to your agent. Paste into your coding agent’s chat. Your agent reviews the source and guides the setup.
Try it with your agent
After installation, start with a request like:
Use the Add Skill skill to help with my task.
Ask me for the context you need before getting started.Requirements & license
- A coding agent that can access your project files and install skills.
- Your agent checks each skill’s instructions, dependencies, and permissions during setup.
- Reported license for the skills shown: MIT. The author’s terms apply.