DeepSeek Harness Plugin

Ri0n72Y/dsh-workspace-scope

Stars ★ 6 Downloads (30d) 1,573 Category UI Enhancements Added 2026-08-18 npm dsh-workspace-scope

Per-workspace Skill and MCP enablement for DeepSeek Harness: control which skills and MCP servers new sessions load, per project.

Install

# from npm (prebuilt)

dsh plugin --profile web add dsh-workspace-scope

# from GitHub (first run asks for allowBuilds approval — follow the hint, retry)

dsh plugin --profile web add github:Ri0n72Y/dsh-workspace-scope

Any plugin you install runs third-party code with your own permissions — it can read your files, use your credentials, and reach the network, and tool approvals don’t sandbox it. GitHub-sourced plugins also run build scripts at install time — pnpm blocks those until you allow them, so an install can stop with ERR_PNPM_GIT_DEP_PREPARE_NOT_ALLOWED or ERR_PNPM_IGNORED_BUILDS; dsh prints the exact key to add under allowBuilds in your profile’s pnpm-workspace.yaml, and the install works on the next run. Allowing a build is a trust decision: only install sources you trust, and pin a commit (github:owner/repo#sha).

README

ci

Enable or disable the current Agent's existing Skills and inherited Host-global MCP servers per DeepSeek Harness workspace.

This plugin only controls capability scope. It does not install, discover, or provide Skills or MCP servers. Each workspace can expose a different capability set through its own .dsh-scope.json.

中文版:README.md

[!IMPORTANT] dsh-workspace-scope v0.5.3 is adapted and tested only with DeepSeek Harness 0.1.6-alpha.2. This does not mean “compatible with 0.1.6-alpha.2 and all later releases.” Future DSH versions may change Client or Host contracts again; validate compatibility before upgrading DSH independently.

Install

Requires the dsh CLI:

dsh plugin --profile web add dsh-workspace-scope

Remove it with:

dsh plugin --profile web remove dsh-workspace-scope

Usage

In a blank new-session Session, click "Workspace scope" on the right side of the input card.

The dialog shows model-invocable Skills from the current Agent and Host-global MCP servers. Switching Agent Presets refreshes the list for the current Agent. Toggles, search, enable-all, and disable-all save immediately to .dsh-scope.json in the workspace root.

The configuration locks when the conversation starts its first real model request. Changes made before that point apply to the conversation; later changes apply only to new conversations.

Configuration

Normally the UI is enough. You can also edit .dsh-scope.json manually:

{
  "default": {
    "mode": "whitelist",
    "skills": ["<skill-name>"],
    "mcps": ["<server-name>"]
  }
}
  • default: enable everything.
  • whitelist: enable only listed capabilities.
  • blacklist: disable listed capabilities.
  • skills: Skill names.
  • mcps: Host-global MCP server names.

Scope

  • Skills come from the current Agent's existing model-invocable Skill view; this plugin does not install or add Skills.
  • Disabling a Skill controls model invocation. Whether /skill-name remains available is determined by that Skill's native DSH userInvocable policy.
  • MCP management covers only Host-global MCP servers inherited by the Agent. Tool / MCP registrations inside Agent or Preset scope are not managed.
  • In 0.5.x, MCP serverName values containing __ are unsupported; ordinary tool names may contain __.

Implementation details, data flow, and C4 diagrams live in docs/architecture.md. Release history lives in CHANGELOG.md.

Contributing

Open an issue for bugs or ideas. Read CONTRIBUTING.md before submitting a pull request.

License

MIT

Content from the project README on GitHub ↗

Links

More in this category

View the whole category →

Community comments

Comments are public GitHub Discussions. Loading them connects to GitHub and Giscus; a GitHub account is required to post.