dbs-chatroom (dbskill): a skill that stages a chat between expert personas

dbs-chatroom makes your agent pick 3 to 5 expert personas for your topic, simulate each one's view, and end each round with a judge's summary.

Installs: 27.2K; on skills.sh
Stars: 11K; on GitHub
Files: 1; besides SKILL.md
Updated: Oct 2026; last repo commit

Give it a topic. If you do not name anyone, the agent proposes 3 to 5 experts with one line each on their framework and role, then stops until you confirm. If you name people, such as /dbs-chatroom 查理·芒格丹尼尔·卡尼曼雷·达里奥, it starts at once. Each expert speaks in about 200 characters, independently of the others, and a judge then rates the discussion, adds blind spots and gives 2 to 3 actions.

The chat stays open for follow-ups until you end it, or type /dbs. The first round opens with a notice that this is a simulation from public writings: quoted words are not the person's own. It runs experts as parallel sub-agents when the host supports them, and one after another when not.

Good for

Founders and creators weighing a decision from several theoretical angles
Students who want contrasting frameworks applied to one question
Anyone who wants disagreements surfaced, not a single answer

Install dbs-chatroom

Claude Code

Copy this prompt.

Install the dbs-chatroom skill in Claude Code. Run this command:

npx skills add https://github.com/dontbesilent2025/dbskill --skill dbs-chatroom

Then check that it loaded and tell me how to use it.

2
Paste it into Claude Code. It shows the change and asks before running it.

Requires Node.js.

From its listing on skills.sh. Source.

What's in the skill

When it loads

根据话题推荐或接受用户指定的专家，模拟多角色对话并总结分歧。用户要求专家讨论、定向聊天室或继续当前讨论时使用。

Sections in SKILL.md

聊天室的边界
工作流程
第一步：检测模式
第二步：推荐人物（仅情况 B）
第三步：生成对话
第四步：判官总结
注意事项

Key facts

Author: dontbesilent (dbskill)
Repo: dontbesilent2025/dbskill, path skills/dbs-chatroom
Installs on skills.sh: 27,244 (as of Nov 2, 2026)
GitHub stars: 11,039 (shared repo, Nov 2, 2026)
License: CC BY-NC 4.0 per the LICENSE file and README badge; GitHub's API shows no recognized license
Last repo commit: Oct 28, 2026
skills.sh audits: Gen Agent Trust Hub: pass. Socket: pass. Snyk: pass.
Files: 2 including SKILL.md: agents/openai.yaml (no scripts)
Language: SKILL.md is written in Chinese

What it includes

/dbs-chatroom [name1] [name2] [name3]
Starts the discussion straight away with the people you chose.
Case A: you name the experts
Recommends 3 to 5 experts and waits
One line per person: framework, role and how far they move the topic. Waits for your confirmation.
Case B: no names given
Picks theories first, people second
Each theory needs a clear claim, a long track record and its own variable. It excludes people known mainly for style or reach.
Recommendation rules
About 200 characters per expert
Independent prompts per expert, run as parallel sub-agents when the host supports them, with a simulation notice on the first round.
Round of expert speeches
Discussion quality, blind spots, 2-3 actions
Closes every round. The chat remains open after it.
Judge's summary
Same experts, new question
Reuses the confirmed people, topic and last summary. Does not re-recommend.
Case C: follow-up input
Say "end chatroom" or type /dbs
Replies that the chat has ended and hands control back to the main agent.
Case D: end the chat

From the official docs

Questions

What does dbs-chatroom do?

It simulates a panel discussion. The agent proposes or accepts 3 to 5 expert personas for your topic, writes a short statement in each one's voice, and then adds a judge's summary of the clash, blind spots and next steps.

How do I install it in Claude Code?

Use the install command in the box above, which installs this skill alone. The dbskill README also lists an install-everything command for all 34 skills across every agent, and a Claude Code plugin.

Are the quotes real?

No. The first round says the text is simulated from public books and statements and is not the person's own words. SKILL.md also tells the agent not to invent quotes or have a person take a stand on a named company, product or living person.

Can I choose the experts myself?

Yes. List them after the command, for example /dbs-chatroom 查理·芒格丹尼尔·卡尼曼雷·达里奥. Named living people are allowed because they are your choice, but SKILL.md says the system does not recommend them by itself.

Is it safe?

SKILL.md bundles no scripts and names no network calls. When your host supports sub-agents, it starts one per expert in parallel. The persona text is generated and is not advice from the real people.

How do I stop it?

Say you want to end the chatroom, or type /dbs. Until then your next messages are treated as follow-ups for the same experts.

Does it need the rest of dbskill?

No. It works alone. SKILL.md mentions the /dbs router only as an optional way back to the main agent.

What is the license?

The LICENSE file and README badge say CC BY-NC 4.0, which bars commercial use. GitHub's API shows no recognized license, so read the LICENSE file before using it at work.

Information

updated Nov 2, 2026

Repository: dontbesilent2025/dbskill
Folder: skills/dbs-chatroom

dontbesilent2025/dbskill skills.sh

Sources

Docs, issue trackers and discussions this page is built from.

Show all 3

SKILL.md · github.com
dbskill README · github.com
dbskill LICENSE · github.com