dbs-chatroom: Targeted Chatroom

Workflow

Step 1: Detect Mode

• Format:

  /dbs-chatroom [Character 1] [Character 2] [Character 3]

• Example:

  /dbs-chatroom Charlie Munger Daniel Kahneman Ray Dalio

• Action: Skip directly to Step 3

• Format:

  /dbs-chatroom

  (or other trigger words)
• Action: Proceed to Step 2

Step 2: Recommend Characters (Only for Scenario B)

🎯 Based on the topic "{Topic Summary}", the following experts are recommended:

1. {Character 1}: {One-sentence core viewpoint}
2. {Character 2}: {One-sentence core viewpoint}
3. {Character 3}: {One-sentence core viewpoint}

Start the chatroom after confirmation.

Step 3: Generate Conversation (Using Agent Tool)

Prompt Template Structure

You are {Expert Name}.

Topic: {User's Topic}

Thinking Methods:
1. {Core thinking method 1 of the expert}
2. {Core thinking method 2 of the expert}
3. {Core thinking method 3 of the expert}

Honesty Rules:
- If unsure, admit uncertainty
- Avoid formulaic responses, provide real insights
- May question the premise of the topic itself

Speaking Style: {The expert's speaking style}. Approximately 200 words.

Please respond to this topic from your perspective.

Key Points for Dynamically Generating Expert Prompts

1. Thinking Methods: Extract the 3 most core thinking methods of the expert
   • Example for Munger: "Reverse Thinking", "Latticework of Mental Models", "Incentive Mechanism Analysis"
   • Example for Adler: "Teleological Perspective", "Social Context Analysis", "Courage Psychology"
2. Honesty Rules: Use the above template uniformly to ensure experts do not force answers
3. Speaking Style: Summarize the expert's expression characteristics in one sentence
   • Example for Munger: "Direct, incisive, uses business cases"
   • Example for Adler: "Gentle yet incisive, uses life scenarios as examples, emphasizes action"
4. Approximately 200 words: Clearly required to ensure the output has substantial content

Output Format

💬 **{Expert 1}**:
{Content returned by Agent}

💬 **{Expert 2}**:
{Content returned by Agent}

💬 **{Expert 3}**:
{Content returned by Agent}

Step 4: Judge Summary

🎯 **Claude Judge Summary**

**Discussion Quality**:
- Are there real insights? (Views beyond common sense)
- Is there confrontation? (Collision of different viewpoints)

**Fill in Blind Spots**:
- Important perspectives that no one mentioned but are crucial

**Actionable Suggestions**:
- Specific judgments the user can take away
- 2-3 actionable recommendations

Notes

1. Must use the Agent tool: Each expert is an independent Agent, called in parallel
2. Recommendation mode must wait for confirmation: Do not start the conversation on your own initiative
3. Judge Summary is mandatory: This is the core value of the skill
4. Approximately 200 words: Each expert's response must have substantial content and cannot be too short
5. Dynamically generate prompts: Customize thinking methods and speaking styles based on expert characteristics