diff --git a/.claude-plugin/marketplace.json b/.claude-plugin/marketplace.json
index 4cd6582..64d0052 100644
--- a/.claude-plugin/marketplace.json
+++ b/.claude-plugin/marketplace.json
@@ -9,7 +9,7 @@
{
"name": "harness",
"source": "./",
- "description": "에이전트 팀 & 스킬 아키텍트. 도메인/프로젝트에 맞는 하네스를 구성하고, 전문 에이전트를 정의하며, 에이전트가 사용할 스킬을 생성하는 메타 스킬.",
+ "description": "에이전트 팀 & 스킬 아키텍트. 도메인/프로젝트에 맞는 하네스를 구성하고, 전문 에이전트를 정의하며, 에이전트가 사용할 스킬을 생성하는 메타 스킬. 团队架构工厂：一句话生成 agent team 与 skills。",
"version": "1.2.0"
}
]
diff --git a/.claude-plugin/plugin.json b/.claude-plugin/plugin.json
index e6d0b9a..3ccd04c 100644
--- a/.claude-plugin/plugin.json
+++ b/.claude-plugin/plugin.json
@@ -1,6 +1,6 @@
{
"name": "harness",
- "description": "The team-architecture factory for Claude Code — a meta-skill that turns a domain description into an agent team and the skills they use, with six pre-defined team-architecture patterns (Pipeline, Fan-out/Fan-in, Expert Pool, Producer-Reviewer, Supervisor, Hierarchical Delegation). Claude Code용 팀 아키텍처 팩토리: 도메인 한 문장을 에이전트 팀과 스킬 세트로 변환하는 메타 스킬.",
+ "description": "The team-architecture factory for Claude Code — a meta-skill that turns a domain description into an agent team and the skills they use, with six pre-defined team-architecture patterns (Pipeline, Fan-out/Fan-in, Expert Pool, Producer-Reviewer, Supervisor, Hierarchical Delegation). Claude Code용 팀 아키텍처 팩토리. Cursor/Claude Code 团队架构工厂：一句话生成 agent team 与 skills。",
"version": "1.2.0",
"author": {
"name": "robin",
diff --git a/CHANGELOG.md b/CHANGELOG.md
index 1be8060..42b3983 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -5,6 +5,9 @@
## [Unreleased]
### Added
+- 简体中文文档：`README_ZH.md`、`docs/quickstart_ZH.md`、`docs/experimental-dependency_ZH.md`
+- SKILL.md / plugin 描述：中文 trigger 短语
+- README EN/KO/JA：语言切换器增加 中文
- 신규 에이전트/스킬 생성 전 중복 검토 단계 (Phase 3-0, Phase 4-0)
- `references/agent-design-patterns.md` "에이전트 재사용 설계" 섹션
- `references/skill-writing-guide.md` §9 "스킬 재사용 설계"
diff --git a/README.md b/README.md
index b64e786..a56e63c 100644
--- a/README.md
+++ b/README.md
@@ -14,12 +14,12 @@
- 
+ 
# Harness — The Team-Architecture Factory for Claude Code
-**English** | [한국어](README_KO.md) | [日本語](README_JA.md)
+**English** | [한국어](README_KO.md) | [日本語](README_JA.md) | [中文](README_ZH.md)
> **Harness is a team-architecture factory for Claude Code.** Say **"build a harness for this project"** (English) or **"하네스 구성해줘"** (한국어) or **"ハーネスを構成して"** (日本語), and the plugin turns your domain description into an agent team and the skills they use — picked from six pre-defined team-architecture patterns.
diff --git a/README_JA.md b/README_JA.md
index 542ce3e..befaeff 100644
--- a/README_JA.md
+++ b/README_JA.md
@@ -14,12 +14,12 @@
-
+
# Harness — Claude Code のためのチームアーキテクチャファクトリー
-[English](README.md) | [한국어](README_KO.md) | **日本語**
+[English](README.md) | [한국어](README_KO.md) | **日本語** | [中文](README_ZH.md)
> **Harness は Claude Code 向けのチームアーキテクチャファクトリーです。** **「ハーネスを構成して」** (日本語) · **"build a harness for this project"** (English) · **"하네스 구성해줘"** (한국어) と伝えるだけで、プラグインがドメイン記述をエージェントチームとそのチームが使うスキルへと変換します — あらかじめ定義された 6 種類のチームアーキテクチャパターンから 1 つを選んで。
diff --git a/README_KO.md b/README_KO.md
index 891994d..7295b8b 100644
--- a/README_KO.md
+++ b/README_KO.md
@@ -14,12 +14,12 @@
-
+
# Harness — Claude Code를 위한 팀 아키텍처 팩토리
-[English](README.md) | **한국어** | [日本語](README_JA.md)
+[English](README.md) | **한국어** | [日本語](README_JA.md) | [中文](README_ZH.md)
> **Harness는 Claude Code용 팀 아키텍처 팩토리입니다.** **"하네스 구성해줘"** (한국어) · **"build a harness for this project"** (English) · **"ハーネスを構成して"** (日本語) 한 문장으로, 플러그인이 도메인 설명을 에이전트 팀과 그들이 쓸 스킬로 변환합니다 — 사전 정의된 6가지 팀 아키텍처 패턴 중 하나를 골라서요.
diff --git a/README_ZH.md b/README_ZH.md
new file mode 100644
index 0000000..e7e9ddc
--- /dev/null
+++ b/README_ZH.md
@@ -0,0 +1,139 @@
+
+ 
+
+
+
+ 
+ 
+ 
+ 
+ 
+
+
+
+
+
+
+
+
+# Harness — Claude Code 的团队架构工厂
+
+[English](README.md) | [한국어](README_KO.md) | [日本語](README_JA.md) | **中文**
+
+> **Harness 是 Claude Code 的团队架构工厂。** 说 **「帮我配置 harness」**（中文）、**「build a harness for this project」**（英文）或 **「ハーネスを構成して」**（日文），插件会根据你的领域描述生成 agent team 及其 skills，并从 6 种预定义团队架构模式中选择。
+
+## 概述
+
+Harness 利用 Claude Code 的 agent team 系统将复杂任务分解为协调工作的专业 agent。说「帮我搭建 harness」即可自动生成 `.claude/agents/` 与 `.claude/skills/`。
+
+## 类别 — Harness 的定位
+
+Harness 位于 Claude Code 生态的 **L3 Meta-Factory** 层 — 生成其他 harness 的工厂。在 L3 内，我们专注 **Team-Architecture Factory** 子层。
+
+| 层级 | 作用 | 相邻项目 |
+|------|------|----------|
+| **L3 — Meta-Factory / Team-Architecture Factory**（本项目） | 领域一句话 → agent team + skills，6 种团队模式 | — |
+| L3 — Runtime-Configuration Factory | 确定性、可重复的运行时配置 | [Archon](https://github.com/coleam00/Archon) |
+| L3 — Codex Runtime Port | 同概念，Codex 运行时 | [meta-harness](https://github.com/SaehwanPark/meta-harness) |
+| L2 — Cross-Harness Workflow | 跨 harness 标准化 skills/rules/hooks | [ECC](https://github.com/affaan-m/everything-claude-code) |
+
+## 核心特性
+
+- **Agent Team 设计** — Pipeline、Fan-out/Fan-in、Expert Pool、Producer-Reviewer、Supervisor、Hierarchical Delegation
+- **Skill 生成** — Progressive Disclosure，高效上下文管理
+- **编排** — agent 间数据传递、错误处理、团队协调协议
+- **验证** — 触发器验证、dry-run、with-skill vs without-skill 对比测试
+
+## 工作流
+
+```
+Phase 1: 领域分析
+ ↓
+Phase 2: 团队架构设计
+ ↓
+Phase 3: Agent 定义 (.claude/agents/)
+ ↓
+Phase 4: Skill 生成 (.claude/skills/)
+ ↓
+Phase 5: 集成与编排
+ ↓
+Phase 6: 验证与测试
+```
+
+## 安装
+
+### Marketplace
+
+```shell
+/plugin marketplace add revfactory/harness
+/plugin install harness@harness-marketplace
+```
+
+### 全局 Skill
+
+```shell
+cp -r skills/harness ~/.claude/skills/harness
+```
+
+## 用法
+
+触发示例：
+
+```
+帮我配置 harness
+搭建 harness：深度研究团队
+设计 agent team 用于多源新闻采集
+Set up a harness for this project
+```
+
+## 架构模式
+
+| 模式 | 说明 |
+|------|------|
+| Pipeline | 顺序依赖任务 |
+| Fan-out/Fan-in | 并行独立任务后合并 |
+| Expert Pool | 按上下文选择性调用专家 |
+| Producer-Reviewer | 生成后质量审查 |
+| Supervisor | 中央 agent 动态分派 |
+| Hierarchical Delegation | 自上而下递归委派 |
+
+## 用例 — 可复制 Prompt
+
+**深度研究**
+```
+Build a harness for deep research. Parallel agents for web, academic,
+and community sources — cross-validate and produce a report.
+```
+
+**多源新闻情报**
+```
+帮我搭建 harness：多源新闻情报。RSS、社交媒体、网页并行采集，
+交叉验证后输出 JSON + markdown 简报。
+```
+
+**代码审查**
+```
+Build a harness for code review — parallel agents for architecture,
+security, performance, style — merge into one report.
+```
+
+## 与相邻项目共存
+
+| 项目 | 定位 | 与 Harness 关系 |
+|------|------|----------------|
+| [Agent-Reach](https://github.com/Panniantong/Agent-Reach) | 互联网接入安装/路由 | **工厂 ↔ 工具供应** — 研究/新闻 harness 的 CLI 层 |
+| [wshobson/agents](https://github.com/wshobson/agents) | agent/skill 目录 | **工厂 ↔ 零件供应** |
+| [LangGraph](https://langchain-ai.github.io/langgraph/) | 状态图编排 | **不同赛道** — 长期有状态编排 |
+
+## 要求
+
+- Agent Teams：`CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1`
+
+## 文档
+
+- [快速上手（中文）](docs/quickstart_ZH.md)
+- [Experimental 依赖（中文）](docs/experimental-dependency_ZH.md)
+
+## License
+
+Apache 2.0
diff --git a/docs/experimental-dependency_ZH.md b/docs/experimental-dependency_ZH.md
new file mode 100644
index 0000000..8047078
--- /dev/null
+++ b/docs/experimental-dependency_ZH.md
@@ -0,0 +1,35 @@
+# Experimental 依赖 — Agent Teams 标志
+
+Harness 当前依赖 Claude Code 的 **Experimental Agent Teams** API。
+
+## 必需环境变量
+
+```bash
+export CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1
+```
+
+未设置时，多 agent 团队可能无法实例化（仅单 agent 响应）。
+
+## 为什么需要
+
+Harness 的默认执行模式使用 `TeamCreate`、`SendMessage`、`TaskCreate`。这些 API 在 Agent Teams experimental 通道下可用。
+
+## 标志变更时
+
+若 Anthropic 将 Agent Teams 提升为 stable：
+
+1. 本仓库 README 将在 **72 小时内** 更新（见 CONTRIBUTING SLA）
+2. 生成的 harness 文件结构不变；仅运行时标志可能移除
+
+## 验证
+
+```bash
+claude --version # 需 2.x+
+claude plugin list # 需 harness@harness
+echo $CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS # 需为 1
+```
+
+## 相关链接
+
+- [Claude Code Agent Teams 文档](https://code.claude.com/docs/en/agent-teams)
+- 英文版：[`experimental-dependency.md`](./experimental-dependency.md)
diff --git a/docs/quickstart_ZH.md b/docs/quickstart_ZH.md
new file mode 100644
index 0000000..0da7775
--- /dev/null
+++ b/docs/quickstart_ZH.md
@@ -0,0 +1,74 @@
+# 快速上手 — 5 分钟生成第一个 Harness
+
+> **时间预算：5 分钟。** 若 5 分钟内无法完成 Step 5，请提 issue — 这是文档问题，不是你的问题。
+
+**完成后你将拥有：** 当前目录下 `.claude/agents/` 中 3–5 个领域专用 agent，以及对应 `.claude/skills/`，由一句话 prompt 生成。
+
+**前置条件：**
+- Claude Code **v2.x+**（`claude --version` 返回 `2.x` 或更高）
+- 支持 `export` 的 shell（bash / zsh / fish）
+- 可访问 `github.com` 与 `api.anthropic.com`
+
+---
+
+## Step 1 — 添加 marketplace（约 60 秒）
+
+```bash
+claude plugin marketplace add revfactory/harness
+```
+
+**预期输出：** `Added marketplace: revfactory/harness`
+
+---
+
+## Step 2 — 安装插件并启用 Experimental 标志（约 40 秒）
+
+```bash
+claude plugin install harness@harness
+export CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1
+```
+
+将 `export` 写入 `~/.zshrc` 或 `~/.bashrc` 以持久化。
+
+详见 [`docs/experimental-dependency_ZH.md`](./experimental-dependency_ZH.md)。
+
+---
+
+## Step 3 — 一句话生成 harness（约 2 分钟）
+
+```bash
+claude "帮我配置 harness：金融科技风险评估团队"
+```
+
+**也可尝试：**
+- `claude "build a harness for a fintech risk-assessment team"`（英文）
+- `claude "하네스 구성해줘 — 핀테크 리스크 평가 팀"`（韩文）
+- `claude "搭建 harness：电商欺诈检测工作流"`
+
+**预期输出：** 流式计划，随后确认已写入 3–5 个 agent 与 skill 文件。
+
+---
+
+## Step 4 — 验证生成文件（约 30 秒）
+
+```bash
+ls -la .claude/agents/
+ls -la .claude/skills/
+```
+
+---
+
+## Step 5 — 用 sample task 跑团队（约 90 秒）
+
+```bash
+claude "Ticket FIN-427: 新客户申请 500 万美元营运资金额度。输出一页 risk memo，含 go/no-go 建议。"
+```
+
+**常见问题：** 新终端未 `export CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1` → 重新 export 后再跑。
+
+---
+
+## 下一步
+
+- [`docs/experimental-dependency_ZH.md`](./experimental-dependency_ZH.md)
+- [harness-100](https://github.com/revfactory/harness-100) — 100+ 预构建 harness 目录
diff --git a/skills/harness/SKILL.md b/skills/harness/SKILL.md
index 83a09c3..6b87484 100644
--- a/skills/harness/SKILL.md
+++ b/skills/harness/SKILL.md
@@ -1,6 +1,6 @@
---
name: harness
-description: "하네스를 구성합니다. 전문 에이전트를 정의하며, 해당 에이전트가 사용할 스킬을 생성하는 메타 스킬. (1) '하네스 구성해줘', '하네스 구축해줘' 요청 시, (2) '하네스 설계', '하네스 엔지니어링' 요청 시, (3) 새로운 도메인/프로젝트에 대한 하네스 기반 자동화 체계를 구축할 때, (4) 하네스 구성을 재구성하거나 확장할 때, (5) '하네스 점검', '하네스 감사', '하네스 현황', '에이전트/스킬 동기화' 등 기존 하네스 운영/유지보수 요청 시 사용."
+description: "하네스를 구성합니다. 전문 에이전트를 정의하며, 해당 에이전트가 사용할 스킬을 생성하는 메타 스킬. (1) '하네스 구성해줘', '하네스 구축해줘' / '帮我配置 harness', '搭建 harness', '构建 harness' 요청 시, (2) '하네스 설계', 'harness 工程化', '设计 harness' 요청 시, (3) 새 도메인/프로젝트 harness 구축, (4) harness 재구성·확장, (5) '하네스 점검'/'harness 审计'/'harness 现状'/'agent/skill 同步' 등 운영·유지보수 요청 시 사용."
---
# Harness — Agent Team & Skill Architect