Visual Agent Design — AI skill for Claude Code
Visual Agent Design (VAD) — A visual-first framework for AI task diagnosis, routing, Visual Agent Cards, agent execution, validation, and human–AI collaboration.
How to install Visual Agent Design
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open draiagent/Visual-Agent-Design and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Visual Agent Design does
Visual Agent Design (VAD) — A visual-first framework for AI task diagnosis, routing, Visual Agent Cards, agent execution, validation, and human–AI collaboration.
Alternatives in AI
- Oh My Claudecode — 32 specialized agents and 7 execution modes for Claude Code, with smart model routing and automatic paralleliz 10.9k ★
- Open Multi Agent — TypeScript multi-agent orchestration engine — one runTeam() call from goal to result 5.8k ★
- Golutra — Multi-agent AI orchestration platform for automation, workflows, and developer tools 3.8k ★
README
Visual Agent Design(VAD)|視覺代理設計
**先判斷任務,再選擇 AI,最後把工作畫給 Agent 看。**
**版本:1.0.0** **語言:繁體中文(zh-TW)** **定位:公開教學/跨模型 Agent Framework/企業 AI 導入/Visual Task Interface/研究方法論**
Visual Agent Design(VAD)不是單一提示詞、單一 Skill,也不是單純的資訊圖卡。它是一套把「任務診斷、智能路由、視覺任務規格、Agent 執行、成果驗收、知識回存」整合在一起的人機協作方法論。
Start Here|第一次使用 VAD
QUICKSTART.md— 10 分鐘開始使用 VADCARD-REGISTRY.md— 標準 Visual Agent Cardsdocs/METHODOLOGY.md— 完整方法論templates/TRC-3D.md— AI Task Routing Cubetemplates/VAC-8.md— Visual Agent Card 八區標準research/RESEARCH-PROTOCOL.md— 可重複研究程序docs/VSC-INTERFACE.md— 讀取 Visual Skill Composer manifest 的入站接口
核心架構
任務需求
↓
TRC-3D 任務三維路由
↓
CARD-REGISTRY 標準任務卡匹配
↓
選擇 Prompt / Research / Monitoring / Workflow / Agent
↓
VAC-8 視覺任務卡
↓
模型 + Skill + Tool + Knowledge + MCP / A2A
↓
Agent 執行
↓
驗收 / Human Review
↓
案例回存與版本優化
三個核心標準
1. TRC-3D|AI Task Routing Cube
- X:任務資訊已知程度:未知 ↔ 已知
- Y:任務發生頻率:單次 ↔ 連續
- Z:任務推理深度:快速處理 ↔ 複雜推理
2. VAC-8|Visual Agent Card
每張可執行視覺任務卡包含八區:
- Task Goal
- Input Assets
- Process Flow
- Tools & Capabilities
- Decision Rules
- Constraints
- Output Specification
- Acceptance Criteria
3. VAD Agent Blueprint
`GOAL | ROLE | SKILLS | TOOLS | KNOWLEDGE | WORKFLOW | DECISION | SUB-AGENTS | MCP/A2A | QA/GOVERNANCE`
**TRC-3D 決定「怎麼做」;VAC-8 定義「這次任務怎麼交付」;VAD Agent Blueprint 定義「Agent 本身怎麼設計」。**
Standard VAC Registry
註冊表分兩層:**Core** 為具名標準 Standard VAC Five-Pack(五張,名稱與內容不變), **Extended** 為後續加入的標準卡(`VAC-INFOGRAPHIC-001` 資訊圖卡、`VAC-COMIC-001` 漫畫筆記)。 兩層在執行、驗證與路由上地位相同,詳見 `CARD-REGISTRY.md`。
Standard VAC Five-Pack
| Card ID | 任務 |
|---|---|
VAC-VIDEO-001 |
影片剪輯 |
VAC-SLIDE-001 |
簡報製作 |
VAC-WEB-001 |
網站生成 |
VAC-DATA-001 |
數據分析 |
VAC-REPORT-001 |
報告製作 |
跨模型使用
本 Repository 提供:
AGENTS.md— OpenAI Codex / Repository Agent 指令入口CHATGPT.md— ChatGPT Project 使用入口CLAUDE.md— Claude Code 使用入口GEMINI.md— Gemini CLI 使用入口- `.agents/skills/visual
Related Skills
Development Analyst
You are an expert in analyzing development sessions and optimizing AI-human collaboration. Your task is to ref
Opuspad
A local-first, WYSIWYG Markdown editor for Chrome. Seamlessly edit local files with a Notion-like experience,
Agentic Engineering
Operate as an agentic engineer using eval-first execution, decomposition, and cost-aware model routing.
Kolbo Skills
Official Kolbo.AI skills package for Claude Code, Cursor, and Codex. 11 self-contained skills: generate, creat
Agent Knowledge System
A knowledge + task system designed for AI consumers, with human-readable docs as a first-class output — Astro-
Dsh Control
Let AI agents operate DeepSeek Harness (DSH) — workspace/session control, task cards, multi-model fanout revie
Related Agents
Task Coordinator
Use this agent when starting ANY task in the UBERON ontology project. This agent MUST be used first and proact
Ff Vision
Flapforge vision and design analyst: reads reference screenshots and writes the canonical visual spec that eve
Ndv Flow
Fleet orchestrator. Use when the work is too large for one agent — PRDs, epics, multi-task workloads, anything