AI Agent Hub
Back to plugins
⚙️

dsh-mcp-servers-panel

Workflow Updated 2026.09.11

Run the following command in DeepSeek Harness:

dsh plugin install whiteS18/dsh-mcp-servers-panel

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

Run dsh plugin install whiteS18/dsh-mcp-servers-panel in the DeepSeek Harness terminal to install this plugin; the full source repository is at https://github.com/whiteS18/dsh-mcp-servers-panel .

About this plugin

Managing MCP servers inside DSH (DeepSeek Harness) has historically meant juggling scattered JSON config files with no visual surface to confirm whether a given server is online, disabled, or erroring. dsh-mcp-servers-panel brings the full MCP server lifecycle into the DSH settings sidebar so you can add, edit, start, stop, and remove servers without leaving the app.\n\nOn launch, an MCP Servers tab appears in the sidebar listing every configured server with its name, scope, and a live status indicator. Clicking a server expands it to reveal the discovered tool count and descriptions. Editing offers two modes: a form mode covering transport protocol, executable command, arguments, environment variables, endpoint URL, and request headers; and a JSON mode that accepts a bare config object, an mcpServers wrapper, or common Cursor / Claude Code formats. Scope can be toggled between user-global (~/.dsh/mcp-servers.json) and workspace (.dsh/mcp.json) at will.\n\nUnder the hood the plugin maintains the MCP process lifecycle over stdio pipes or HTTP, performs the full 2024-11-05 protocol handshake (initialize, notifications/initialized, tools/list), and registers every discovered tool under a stable mcp____ name into ctx.tools, so the model can call them directly in conversation with no bridge scripts.\n\nIdeal for developers running multiple MCP servers in DSH who want a unified native-UI panel for configuration management and immediate model-invoked tool chains, especially when maintaining both user-level and project-level MCP configurations.

Use Cases

  • Invoking tools from multiple MCP servers within one DSH conversation
  • Toggling a specific MCP server on or off without affecting others
  • Maintaining dual user-level and project-level MCP configurations

Best For

  • Developers running multiple MCP servers inside DSH
  • Users who prefer a native UI over hand-editing JSON for MCP config
  • AI workflow practitioners needing direct model access to MCP tools