AI Agent Hub
Back to plugins
🖥️

dsh-group-chat-view

Client Updated 2026.08.22

Run the following command in DeepSeek Harness:

dsh plugin install tang-zhilei/dsh-group-chat-view

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

Run dsh plugin install tang-zhilei/dsh-group-chat-view in your terminal to install the plugin. The source code is available at https://github.com/tang-zhilei/dsh-group-chat-view . After installation, restart DSH and click the Group Chat tab at the top of the session view.

About this plugin

DSH's default conversation view is a flat, linear message stream. When a single turn interweaves system notices, assistant replies, tool calls, reasoning traces, and sub-Agent tasks, the hierarchy quickly becomes cluttered and hard to scan. dsh-group-chat-view remaps the flow into a familiar group-chat interface: every role gets its own avatar, nickname, and bubble color, and messages sit on alternating left-right lanes just like a real group conversation. Tool-call inputs and outputs merge into a single collapsible card so large JSON dumps no longer flood the screen; sub-Agent live streams nest as their own mini group-chat threads you can expand inline; and clicking any avatar pops up a metadata panel with model name, temperature, token usage, and latency. The view also renders full Markdown, follows DSH's dark/light theme, ships Chinese and English localization, and features smart auto-scroll with a back-to-bottom button.

If you rely on tool chains or multi-Agent orchestration in DSH on a daily basis and find the default flat list hard to navigate at a glance, this plugin turns every session into a well-structured group chat. Who said what, which tools were invoked, and how each sub-task progressed all become instantly readable.

Use Cases

  • Collapsing and expanding tool-call inputs and outputs to quickly locate key information in long DSH sessions.
  • Tracking sub-Agent real-time message streams inline without switching windows during multi-Agent orchestration.
  • Inspecting per-message metadata such as model name, token usage, and latency during daily debugging.

Best For

  • Developers and researchers who heavily rely on DSH tool chains.
  • Orchestrators managing multiple sub-Agent tasks concurrently.
  • DSH users who prefer a familiar WeChat-style chat interface for quick scanning.