Back to directory

Awesome DeepSeek Harness (DSH) Plugin

A curated, client-agnostic list of community plugins for DeepSeek Harness, an open-source coding-agent framework where every part of the agent is a plugin.

Awesome DeepSeek Harness (DSH) Plugin

This is a curated list of plugins for DeepSeek Harness (dsh), DeepSeek's open-source agent harness, a runnable coding agent, available in web and headless form, built on a framework where models, tools, sandboxes, session storage, the UI and even the agent loop itself are all plugins. Listed plugins can extend the official coding agent, swap out its core parts, or assemble something different, and each declares a dsh.bundle manifest so it installs through the dsh plugin add command.

The list is explicitly client-agnostic: a plugin is included because it follows the official plugin protocol, not because it was built for a particular desktop client, though several third-party clients bundle a plugin market that makes every listed plugin one click away to install. An entry is added once the plugin installs correctly, does what its one-line description claims, sits in the right category, and is actively maintained; each submission's claims are checked against its source before merging.

The project warns that installing a plugin runs third-party code with the user's own permissions, so being listed is not a security review.

More in Agent skills

See all 581
  • Provenance gate

    github.com

    A deterministic authorization gateway that sits between an AI agent and its tools, tracing where each parameter of a side-effect call came from, and blocking the call or asking a human when the source does not check out.

  • A markdown framework of system rules that restates Isaac Asimov's Four Laws of Robotics as a set of prioritized loss calculations for large-language-model agents to follow, published as a discovery file for agents and crawlers to read.

  • Vitruvius

    github.com

    An engineering research agent that runs a discover, read, synthesize, verify and review loop over technical questions, reading sources directly rather than inferring from titles, and publishing its own scored verifier benchmark, weaknesses included.