Token导航 LogoToken导航

Agent Skill · 研究检索

system-design(系统设计)

用于辅助界面设计、视觉规范、排版、配色、布局和交互体验优化。它适合让 Agent 根据产品场景整理页面结构、生成 UI 方案、检查视觉一致性或改进组件层级。使用时需要结合现有品牌、设计系统和用户任务,不应只堆装饰元素;涉及真实页面改动时,应通过截图或浏览器预览检查文本溢出、对齐和响应式表现。

system-design 是研究检索类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为需要联网;许可证信息还不明确,第三方扫描有提醒项,按安装量在研究检索类的 73,675 个 Skill 里排第 39,123。

总安装 1,458周安装 59GitHub Stars 68下载 458核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
unknown

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:system-design(系统设计)
来源仓库:https://github.com/luongnv89/skills
仓库路径:skills/system-design
安装命令:
npx skills add https://github.com/luongnv89/skills --skill system-design
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/luongnv89/skills --skill system-design

这个 Skill 是做什么的

研究检索类 Skill 主要解决「找资料、核对事实、把来源整理清楚」的问题,通常会让助手先检索再总结。

system-design 用于辅助界面设计、视觉规范、排版、配色、布局和交互体验优化。适合让 Agent 根据产品场景整理页面结构、生成 UI 方案或检查视觉一致性。使用时需要结合现有品牌、设计系统和用户任务,避免堆砌装饰元素。涉及真实页面改动时,应通过截图或浏览器预览检查文本溢出、对齐和响应式表现。

适用宿主包括 Codex、Claude、Cursor、Gemini CLI,接入前应确认版本、权限和运行环境要求。

来源仓库内路径:skills/system-design

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/luongnv89/skills --skill system-design;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它需要需要联网,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

system-design 怎么安装?

在终端执行 npx skills add https://github.com/luongnv89/skills --skill system-design。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

system-design 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

system-design 需要什么权限?

权限级别是需要联网。该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

system-design 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

system-design 用的人多吗?

公开数据是累计安装 1,458 次,近一周安装 59 次,下载 458 次,GitHub 68 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

system-design 安全吗?

第三方扫描有提醒项,说明存在需要人工确认的行为。安装前请打开源码看清它读了什么、联网到哪里。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。

先想清楚的地方

  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 第三方扫描存在提醒项或未通过项,需要人工逐条确认。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理研究检索类任务、又不想每次都手写提示词的人。

不太适合

  • 需要明确授权才能用到公司项目里的场景。
  • 需要厂商级技术支持和 SLA 的生产系统。

研究检索类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
find-skills

寻找技能 · Codex、Claude

31,199,81316,486操作浏览器MIT有提醒
self-improving-agent

自我改进 Agent · OpenClaw

16,464,0023,396接触敏感数据MIT-0通过
soultrace

灵魂追踪 · Codex、Claude

7,588,94412需要联网未明确有提醒
self-improving

自我完善 · OpenClaw

6,829,2841,046需要联网MIT-0通过
microsoft-foundry

微软代工厂 · Codex、Claude

6,614,144764调用外部服务未明确有提醒
system-design

系统设计 · 本页

1,45868需要联网未明确有提醒

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

System Design

Generate comprehensive Technical Architecture Documents with modular design for startups.

Subagent Architecture

This skill uses parallel research agents with upfront content extraction. Pattern: D (Research+Synthesis) + E (Staged Pipeline).

Agents

AgentRoleParallelization
prd-readerRead PRD + supporting docs, return structured extractionSequential (only once)
tech-researcherHandle one research round (spawned 5x in parallel)Parallel (5 instances)
tad-writerGenerate complete tad.md from all inputsSequential (after all research)

Research Rounds (5 Parallel)

  • Round 1: Technology Stack validation (React, Node.js, PostgreSQL, Elasticsearch)
  • Round 2: Infrastructure validation (Vercel, AWS, CDN, cost estimation)
  • Round 3: Security review (auth, encryption, compliance, API security)
  • Round 4: Risk assessment (bottlenecks, vendor lock-in, team gaps)
  • Round 5: Holistic review (PRD alignment, team capability, quick wins)

Parallelization Strategy

  • PRD Content: Extracted once by prd-reader, stays out of main context
  • Research Independence: Each round researches conceptually different angle (tech vs. infra vs. security)
  • Reasoning Isolation: Parallel rounds keep each area's reasoning isolated, prevent groupthink
  • Note: Research rounds are conceptual reasoning, not data fetching — parallel rounds won't produce fundamentally different info, but isolation improves quality

Result: All 5 rounds complete concurrently, tad-writer synthesizes outputs into unified TAD.

Environment Check

Before executing:

  1. Verify prd.md exists in project directory
  2. Check for supporting docs (idea.md, validate.md) if available
  3. Confirm WebSearch and WebFetch tools available for research
  4. Verify write permissions to project root for tad.md creation
  5. Ensure git access for final commit

Repo Sync Before Edits (mandatory)

Before creating/updating/deleting files in an existing repository, sync the current branch with remote:

branch="$(git rev-parse --abbrev-ref HEAD)"
git fetch origin
git pull --rebase origin "$branch"

If the working tree is not clean, stash first, sync, then restore:

git stash push -u -m "pre-sync"
branch="$(git rev-parse --abbrev-ref HEAD)"
git fetch origin && git pull --rebase origin "$branch"
git stash pop

If origin is missing, pull is unavailable, or rebase/stash conflicts occur, stop and ask the user before continuing.

Input

Project folder path in $ARGUMENTS containing:

  • prd.md - Product requirements (required)
  • idea.md, validate.md - Additional context (optional)

Workflow

Phase 1: Setup & Validation

  1. Verify prd.md exists
  2. Read supporting docs if present
  3. Read references/tech-stack.md for technology recommendations
  4. Backup existing tad.md if present

Phase 2: Extract Context

From PRD extract:

  • Product name and vision
  • Core features and requirements
  • User flows
  • Non-functional requirements
  • Third-party integrations
  • Analytics requirements

Phase 3: Clarify Architecture

Ask user (if not clear):

DecisionOptions
DeploymentVercel/Netlify (recommended), AWS, GCP, Self-hosted
DatabasePostgreSQL, MongoDB, Supabase/Firebase, Multiple
AuthSocial (OAuth), Email/password, Magic links, Enterprise SSO
BudgetFree tier, <$50/mo, <$200/mo, Flexible

Phase 4: Research & Validation

Conduct 5 research rounds:

  1. Technology Stack: Validate choices against industry standards
  2. Infrastructure: Compare hosting for cost and scalability
  3. Security: Review OWASP guidelines for chosen stack
  4. Risk Assessment: Identify bottlenecks, vendor lock-in
  5. Holistic Review: Ensure PRD alignment and startup feasibility

Phase 5: Generate TAD

Create tad.md with sections:

  1. System Overview - Purpose, scope, PRD alignment
  2. Architecture Diagram - Mermaid diagrams for system and flows
  3. Technology Stack - Frontend, backend, database, infrastructure, DevOps
  4. System Components - Modular design with interfaces and dependencies
  5. Data Architecture - Schema, models, flows, storage
  6. Infrastructure - Hosting, environments, scaling, CI/CD, monitoring
  7. Security - Auth, authorization, data protection, API security
  8. Performance - Targets, optimization strategies, caching
  9. Development - Environment setup, project structure, testing, deployment
  10. Risks - Risk matrix with mitigations
  11. Appendix - Research insights, alternatives, costs, glossary

See references/tad-template.md for full template structure.

Phase 6: README Maintenance (ideas repo)

After writing tad.md, if the project folder is inside an ideas repo, update the repo README ideas table:

  • Preferred: cd to repo root and run python3 scripts/update_readme_ideas_index.py (if it exists)
  • Fallback: update README.md manually (ensure TAD status becomes ✅ for that idea)

Phase 7: Commit and push (mandatory)

  • Commit immediately after updates.
  • Push immediately to remote.
  • If push is rejected: git fetch origin && git rebase origin/main && git push.

Do not ask for additional push permission once this skill is invoked.

Phase 8: Output

  1. Write tad.md to project folder
  2. Summarize architecture decisions
  3. Highlight modular design benefits
  4. List cost estimates by phase
  5. Suggest next steps (setup dev environment, create tasks)

Reporting with GitHub links (mandatory)

When reporting completion, include:

  • GitHub link to tad.md
  • GitHub link to README.md when it was updated
  • Commit hash

Link format (derive <owner>/<repo> from git remote get-url origin):

  • https://github.com/<owner>/<repo>/blob/main/<relative-path>

Step Completion Reports

After completing each major step, output a status report in this format:

◆ [Step Name] ([step N of M] — [context])
··································································
  [Check 1]:          √ pass
  [Check 2]:          √ pass (note if relevant)
  [Check 3]:          × fail — [reason]
  [Check 4]:          √ pass
  [Criteria]:         √ N/M met
  ____________________________
  Result:             PASS | FAIL | PARTIAL

Adapt the check names to match what the step actually validates. Use √ for pass, × for fail, and — to add brief context. The "Criteria" line summarizes how many acceptance criteria were met. The "Result" line gives the overall verdict.

Phase-specific checks

Phase 1 — Setup

◆ Setup (step 1 of 8 — environment validation)
··································································
  PRD found:                    √ pass
  Context extracted:            √ pass (product + features + NFRs read)
  Architecture questions answered: √ pass (deployment, DB, auth, budget confirmed)
  ____________________________
  Result:             PASS | FAIL | PARTIAL

Phase 4 — Research

◆ Research (step 4 of 8 — validation rounds)
··································································
  5 parallel rounds completed:  √ pass
  Best practices gathered:      √ pass (tech, infra, security, risk, holistic)
  Patterns validated:           √ pass (OWASP, vendor lock-in, startup feasibility)
  ____________________________
  Result:             PASS | FAIL | PARTIAL

Phase 5 — Generation

◆ Generation (step 5 of 8 — TAD authoring)
··································································
  11 sections written:          √ pass
  tad.md created:               √ pass
  Diagrams included:            √ pass (mermaid architecture + flow diagrams)
  ____________________________
  Result:             PASS | FAIL | PARTIAL

Phase 8 — Output

◆ Output (step 8 of 8 — delivery)
··································································
  Summary presented:            √ pass (architecture decisions highlighted)
  README updated:               √ pass (TAD status ✅)
  Committed and pushed:         √ pass (commit hash: ...)
  ____________________________
  Result:             PASS | FAIL | PARTIAL

Modification Mode

For existing TAD changes:

  1. Create timestamped backup
  2. Ask what to modify (stack, infrastructure, security, data, scaling)
  3. Apply changes preserving structure
  4. Update revision history

Guidelines

  • Practical: Implementable solutions for startups
  • Cost-conscious: Consider budget implications
  • Modular: Emphasize separation of concerns
  • Specific: Concrete technology choices
  • Visual: Include mermaid diagrams

平台分布

Codex

33.52%
按下载量换算154

Claude

29.06%
按下载量换算133

Cursor

19.31%
按下载量换算88

Gemini CLI

10.54%
按下载量换算48

安全审计

Gen Agent Trust Hub

通过

Socket

可疑

Snyk

可疑

权限和风险

需要联网

该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。来源安全扫描存在提醒或未通过项,安装前请自行复核。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills