AdYour product here, seen by people installing AI tools every day Sponsor slots open: 12 of 12, from $249 a month

How to install SandBase Harness in VS Code

To install SandBase Harness in VS Code, click "Add to VS Code" on this page or add the "sandbase-harness" block under "servers" in .vscode/mcp.json, then pick it in Copilot Chat agent mode.

Built from the published package · server updated 4 days ago

  1. 1

    One click (fastest)

    Click the button to open VS Code with this server ready to install.

    Add SandBase Harness to VS Code
  2. 2

    Or add it by hand

    Create .vscode/mcp.json and paste this. Note that VS Code uses "servers", not "mcpServers".

    {
      "servers": {
        "sandbase-harness": {
          "command": "docker",
          "args": [
            "run",
            "-i",
            "--rm",
            "-e",
            "MANAGED_AGENTS_URL",
            "ghcr.io/sandbaseai/sandbase-harness-mcp:0.3.8"
          ],
          "env": {
            "MANAGED_AGENTS_URL": "YOUR_MANAGED_AGENTS_URL"
          }
        }
      }
    }
  3. 3

    Start it in Copilot Chat

    Open Copilot Chat in Agent mode and pick the server from the tools menu.

Config file: .vscode/mcp.json in your workspace

MCP servers run with your permissions. If SandBase Harness needs an API key, add it under env in the config. Check its README.

Install SandBase Harness in another app

Similar servers

Questions

Does SandBase Harness work in VS Code?

Yes. VS Code supports local and remote mcp servers in copilot agent mode, and SandBase Harness installs with the steps above.

SandBase Harness does not show up in VS Code. What now?

Restart VS Code fully, check the JSON is valid, and make sure Node.js (for npx) is installed. See common MCP errors and fixes.

Is SandBase Harness safe?

It scores 86 of 100. Review the code before you install it, and only give it the access it needs. Grades and details.