AI Agent Hub
Back to plugins
🖥️

dsh-terminal-tabs

Client Updated 2026.08.26

Run the following command in DeepSeek Harness:

dsh plugin install bbboy31/dsh-terminal-tabs

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

Install the plugin in DeepSeek Harness by running dsh plugin install bbboy31/dsh-terminal-tabs; the full source code lives at https://github.com/bbboy31/dsh-terminal-tabs. Refresh the web page after installation and the new Terminals tab will appear.

About this plugin

Whenever a model kicks off background commands during a session—npm run dev, git operations, subagent tasks—there is no visible entry point in the DeepSeek Harness web UI to inspect what those processes are doing, how long they have been running, or to terminate them. They simply run silently in the blind.

dsh-terminal-tabs registers a native Terminals view tab alongside Chat and Trajectory, listing every background job in the current session as a row. Each row carries a short Chinese purpose title, a collapsible full command line, a per-second elapsed timer, a live scrollback output pane, and a Close button that genuinely kills the process through the node-side HTTP route. The tab label updates in real time to reflect how many jobs are active, giving you an at-a-glance count. The plugin mirrors output at the node level without altering the model-side read contract, so inference is untouched.

It is well suited to developers and AI engineers who regularly let models spin up long-running dev servers, bulk git operations, or subagent tasks inside DeepSeek Harness. Installation is instant—just refresh the page after installing, no build step required. Uninstalling removes the routes, tab, and stylesheets cleanly with no residue.

Use Cases

  • Inspect background commands started by the model with elapsed time
  • Stream live command output and scroll through history
  • Terminate unwanted background jobs with one click

Best For

  • Developers who frequently let models spawn background services
  • Users running subagent tasks in DeepSeek Harness
  • Engineers who need visual oversight of AI session background activity