hs3180/daily-soul-question

Daily Soul Question

Analyze chat/work records and generate thought-provoking questions for group discussion.

When to Use This Skill

Use this skill for:

• Generating daily reflection questions from chat history
• Creating discussion topics for BBS-style topic groups
• Identifying interesting decisions, potential improvements, and reflection-worthy issues

Keywords that trigger this skill: "灵魂拷问", "每日反思", "话题讨论", "soul question", "daily reflection", "生成话题"

Core Principle

Use prompt-based analysis, NOT complex program modules.

The LLM should analyze chat history directly and generate thoughtful questions, not through pre-built algorithms.

Context Variables

When invoked, you receive:

• Chat ID: Feishu chat ID (from "Chat ID: xxx")
• Message ID: Message ID (from "Message ID: xxx")
• Sender Open ID: Sender's open ID (from "Sender Open ID: xxx")

---

Analysis Process

Step 1: Read Chat History

Read the chat history files from workspace/chat/ directory:

1. List all chat history files:

ls -la workspace/chat/*.md

2. Read recent messages (last 7 days) from relevant chats:

# For each relevant chat, read the history
cat workspace/chat/{chatId}.md | tail -500

Step 2: Analyze Records (Prompt-Based)

Analyze the chat/work records to identify:

1. Interesting Decisions: Notable choices made during discussions

   • Technical decisions (architecture, tool choices)
   • Process decisions (workflow changes, priorities)
   • Any decision that sparked debate or consideration

2. Potential Improvements: Areas that could be enhanced

   • Recurring issues or pain points
   • Inefficiencies noticed
   • Technical debt mentioned

3. Reflection-Worthy Issues: Topics that deserve deeper thinking

   • Unanswered questions
   • Trade-offs that weren't fully explored
   • Assumptions that might need validation

Step 3: Generate Soul Question

Create a thought-provoking question that:

• Opens discussion (not yes/no question)
• Connects to real work (based on actual records)
• Invites diverse perspectives (no single right answer)
• Is relevant today (recent context)

Question Template:

## 🤔 今日灵魂拷问

分析今天的聊天/工作记录，发现一个有趣的问题：

「{核心问题}」

{背景说明}

{引导思考的问题}

欢迎在群里讨论 👇

Step 4: Send to Topic Group

Use the send_user_feedback MCP tool to send the question:

send_user_feedback({
  chatId: "{target_chat_id}",
  message: "{generated_question}"
})

Note: The target chat ID should be configured in the schedule file or use the current chat.

---

Example Analysis

Input (Chat History Excerpt):

## [2024-01-15T10:00:00Z] 📥 User
我们在处理 #123 时选择了方案 A 而非方案 B，因为方案 A 更简单

## [2024-01-15T10:05:00Z] 📥 User
但方案 B 的扩展性更好，只是需要更多时间

## [2024-01-15T10:10:00Z] 📤 Bot
理解，这是一个典型的权衡：简单性 vs 扩展性

Output (Soul Question):

## 🤔 今日灵魂拷问

分析今天的聊天/工作记录，发现一个有趣的问题：

「在处理 #123 时，我们选择了方案 A（简单）而非方案 B（扩展性好）。这个决策是否正确？」

背景：
- 方案 A 更简单，可以快速实现
- 方案 B 扩展性更好，但需要更多时间

值得思考：
1. 我们是否低估了未来的扩展需求？
2. "简单优先"在什么情况下是正确的策略？
3. 如果重来一次，你会做同样的选择吗？

欢迎在群里讨论 👇

---

Quality Guidelines

Good Soul Questions:

• ✅ Based on actual work/decisions
• ✅ Open-ended (multiple valid perspectives)
• ✅ Thought-provoking (makes people think)
• ✅ Relevant to the team

Avoid:

• ❌ Yes/no questions
• ❌ Questions with obvious answers
• ❌ Questions that criticize specific people
• ❌ Questions unrelated to actual work

---

Schedule Configuration

To enable daily soul questions, create a schedule file:

---
name: "每日灵魂拷问"
cron: "0 10 * * *"  # Every day at 10:00 AM
enabled: true
blocking: true
chatId: "{your_topic_group_chat_id}"
---

请使用 daily-soul-question skill 分析最近的聊天记录，生成一个灵魂拷问并发送到当前群聊。

要求：
1. 读取 workspace/chat/ 目录下的聊天记录
2. 重点关注最近的决策、讨论和问题
3. 生成一个开放式的、引人思考的问题
4. 使用 send_user_feedback 发送到当前 chatId

---

Checklist

• [ ] Read chat history from workspace/chat/
• [ ] Identified interesting decisions or issues
• [ ] Generated open-ended, thought-provoking question
• [ ] Question is based on actual work context
• [ ] Ready to send using send_user_feedback

---

DO NOT

• Generate questions that criticize individuals
• Create questions unrelated to actual work
• Ask yes/no questions
• Skip the analysis and generate generic questions
• Send multiple questions (one per day is enough)