AI Agent Hub
Back to plugins
🖥️

DSH-pipeline-kernel

Client Updated 2026.08.26

Run the following command in DeepSeek Harness:

dsh plugin install not-big-dog/DSH-pipeline-kernel

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

In DeepSeek Harness, add the source URL https://github.com/not-big-dog/DSH-pipeline-kernel in the plugin installer to install this plugin.

About this plugin

DSH-pipeline-kernel addresses a central problem in multi-agent work: how to manage pipelines without baking business logic into the kernel. It treats a pipeline as configuration data rather than a fixed workflow. You define roles, chains, and gates, while the kernel handles deployment, registration, task submission, claiming, and automatic routing after completion. Concrete capabilities such as image generation or video production stay inside your pipeline configuration, keeping the kernel general-purpose.

Its core strengths are the task board, ledger, deployment and registry, packaging and reuse, control panel, and watchdog. Through pipeline_* tools, you can define and deploy pipelines, seed role sessions, push tasks, inspect overall status, and let completed work move automatically to the next link. The Web control panel shows activity summaries, pipeline graphs, task inspection, creation, and archiving, while the default watchdog periodically re-wakes open tasks, heals zombie routing states, and reports stalled chains to the supervisor.

It is a good fit for developers and teams building multi-agent automation inside DeepSeek Harness, especially those who want collaborative workflows to become definable, observable, and reusable assets. Because no pipelines are preset by default, it suits users who prefer to design their own orchestration rather than inherit fixed templates.

Use Cases

  • Define and deploy custom multi-agent pipelines.
  • Push, claim, and complete tasks with automatic routing.
  • Monitor activity, archives, and watchdog status in the control panel.

Best For

  • Developers orchestrating multi-agent workflows in DSH.
  • Teams turning pipelines into reusable configurations.
  • Automation practitioners needing task ledgers, status panels, and watchdogs.