AI Agent Hub
Back to plugins
dsh-council preview

dsh-council

Client Updated 2026.08.26

Run the following command in DeepSeek Harness:

dsh plugin install fufengyuan/dsh-council

Paste the following prompt into your AI chat to install this plugin:

Run dsh plugin install fufengyuan/dsh-council in DeepSeek Harness to install this plugin; the source code lives at https://github.com/fufengyuan/dsh-council. Restart the dsh web service after installation.

About this plugin

A single model answer to a complex problem almost always reflects one perspective. dsh-council assembles a Council of 18 cross-disciplinary figures inside DeepSeek Harness-Socrates, Sun Tzu, Feynman, Munger, Taleb, Kaiming He, and others-who debate the same question from different angles in a structured multi-round format, with a chair delivering a synthesized verdict that is far richer than a one-shot reply.

The right-hand DAG panel is the centerpiece: topics, each member's contribution, and the chair's ruling appear as live nodes and edges. Click any node to jump straight into that sub-conversation. Five debate modes are available-sprint, duo, triad, full council, and custom member selection-and a project-level .council.yaml pins the default panel and chair. Every session is persisted in a local SQLite database, so history survives restarts; a coordinator reports per-member progress over HTTP to keep the DAG updating in real time. Member toggles, model mappings, and other settings live in a local JSON file, so the plugin works out of the box with zero external services.

It is built for developers and researchers in the dsh ecosystem who need structured, multi-perspective reasoning on architecture reviews, product decisions, technology choices, or trade-off analyses. When a single answer is not enough and you want a moderated round-table with counter-arguments and a final ruling, the council button at the bottom of the left sidebar is all you need.

Screenshots

Use Cases

  • Multi-perspective review and counter-argument of complex architecture proposals
  • Structured trade-off analysis for product decisions and technology selection
  • Deep reasoning requiring multi-round rebuttals and a chaired verdict

Best For

  • Developers in the DeepSeek Harness ecosystem tackling complex decisions
  • Researchers who need multi-angle argumentation instead of a single answer
  • Tech leads conducting technology selection and trade-off evaluation