From 013cc823b6133bcf4f83905de698df0e7c9f7d57 Mon Sep 17 00:00:00 2001
From: jilinchen <jilin.chen@suwen.ai>
Date: Wed, 7 Jan 2026 02:03:28 +0800
Subject: [PATCH] docs: add real workflow examples from AI Agent Demo research
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

- Show actual outline.yaml, JSON output, and report
- Keep examples concise and elegant

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
---
 README.md    | 50 +++++++++++++++++++++++++++++++-------------------
 README.zh.md | 50 +++++++++++++++++++++++++++++++-------------------
 2 files changed, 62 insertions(+), 38 deletions(-)

diff --git a/README.md b/README.md
index 2fa8b4b..f974d62 100644
--- a/README.md
+++ b/README.md
@@ -35,40 +35,52 @@ cp agents/web-search-agent.md ~/.claude/agents/
 | `run /research/deep` | Deep research each item with parallel agents |
 | `run /research/report` | Generate markdown report from JSON results |
 
-## Workflow
+## Workflow & Example
+
+> Example: Researching "AI Agent Demo 2025"
 
 ### Phase 1: Generate Outline
 ```
-run /research <topic>
+run /research AI Agent Demo 2025
+```
+**Output:** `ai-agent-demo/outline.yaml`
+```yaml
+topic: "AI Agent Demo & Review (2025.9-2025.12)"
+items:
+  - name: "ChatGPT Agent"
+    category: "Browser Agent"
+    brief: "OpenAI unified Agent, released July 2025"
+  - name: "Claude Computer Use"
+    category: "Desktop Agent"
+    brief: "Anthropic desktop control Agent"
+  # ... 15 more items
 ```
-- Model knowledge generates initial items and field framework
-- Web search supplements latest items
-- User confirms and adjusts
-- Outputs: `outline.yaml` (items + config) + `fields.yaml` (field definitions)
 
 ### Phase 2: Deep Research
 ```
 run /research/deep
 ```
-- Parallel agents research each item (batch_size configurable)
-- Each agent reads fields.yaml and outputs structured JSON
-- Supports checkpoint resume
-- Outputs: `results/*.json`
-
-### Optional: Expand Outline
-```
-run /research/add-items    # Add research targets via user input or web search
-run /research/add-fields   # Add field definitions
+**Output:** `ai-agent-demo/results/ChatGPT_Agent.json`
+```json
+{
+  "basic_info": {
+    "name": "ChatGPT Agent",
+    "company": "OpenAI",
+    "release_date": "2025-07-17",
+    "pricing": "Pro $200/mo, Plus $20/mo"
+  },
+  "tech_specs": {
+    "underlying_model": "GPT-5 series",
+    "agent_type": "Unified autonomous Agent"
+  }
+}
 ```
 
 ### Phase 3: Generate Report
 ```
 run /research/report
 ```
-- Generates Python script to convert JSON to markdown
-- User selects summary fields for TOC
-- Skips uncertain values automatically
-- Outputs: `report.md`
+**Output:** `ai-agent-demo/report.md` - Markdown report with TOC and all items
 
 ## References
 
diff --git a/README.zh.md b/README.zh.md
index 3cf810f..bdd7b96 100644
--- a/README.zh.md
+++ b/README.zh.md
@@ -35,40 +35,52 @@ cp agents/web-search-agent.md ~/.claude/agents/
 | `run /research/deep` | 使用并行agents对每个item进行深度调研 |
 | `run /research/report` | 从JSON结果生成markdown报告 |
 
-## 工作流
+## 工作流 & 示例
+
+> 示例：调研 "AI Agent Demo 2025"
 
 ### 阶段1：生成Outline
 ```
-run /research <topic>
+run /research AI Agent Demo 2025
+```
+**输出：** `ai-agent-demo/outline.yaml`
+```yaml
+topic: "AI Agent Demo & 测评 (2025.9-2025.12)"
+items:
+  - name: "ChatGPT Agent"
+    category: "浏览器Agent"
+    brief: "OpenAI统一Agent，2025年7月发布"
+  - name: "Claude Computer Use"
+    category: "桌面Agent"
+    brief: "Anthropic桌面操控Agent"
+  # ... 另外15个items
 ```
-- 模型知识生成初始items和字段框架
-- 网络搜索补充最新items
-- 用户确认并调整
-- 输出：`outline.yaml`（items + 配置）+ `fields.yaml`（字段定义）
 
 ### 阶段2：深度调研
 ```
 run /research/deep
 ```
-- 并行agents调研每个item（batch_size可配置）
-- 每个agent读取fields.yaml并输出结构化JSON
-- 支持断点续传
-- 输出：`results/*.json`
-
-### 可选：扩展Outline
-```
-run /research/add-items    # 通过用户输入或网络搜索添加调研对象
-run /research/add-fields   # 添加字段定义
+**输出：** `ai-agent-demo/results/ChatGPT_Agent.json`
+```json
+{
+  "basic_info": {
+    "name": "ChatGPT Agent",
+    "company": "OpenAI",
+    "release_date": "2025-07-17",
+    "pricing": "Pro $200/月, Plus $20/月"
+  },
+  "tech_specs": {
+    "underlying_model": "GPT-5系列",
+    "agent_type": "统一型自主Agent"
+  }
+}
 ```
 
 ### 阶段3：生成报告
 ```
 run /research/report
 ```
-- 生成Python脚本将JSON转换为markdown
-- 用户选择目录中显示的摘要字段
-- 自动跳过不确定值
-- 输出：`report.md`
+**输出：** `ai-agent-demo/report.md` - 带目录和所有item的Markdown报告
 
 ## 参考文献