agentleFS
lintendolintendo/Axhub-Skills/skills/mcp-installer/SKILL.md

mcp-installer

Skill145 starsChanged 19 days agoMITView on GitHub

Install and configure MCP servers across desktop and CLI clients (Claude, Cline, Windsurf, Cursor, VSCode, Gemini CLI, Codex, Trae, Antigravity, etc.) on macOS/Windows/Linux, preferring @smithery/cli when supported and otherwise performing manual JSON config updates and path discovery.

1 other project uses a version of this.

Adopted by

Most recently taken up by

  • VodouAI/OStheir own version, first seen Jul 19, 2026

The file

We split this skill into the rules it gives an agent

8 rules. About 0 are common, 4 are distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈4

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 4

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Run npx @smithery/cli --help and npx @smithery/cli install --help to confirm the exact syntax.

    line 22
  2. Distinctive:

    Use the CLI install flow for the client ID.

    line 23
  3. Very specific:

    Read existing JSON and ensure a top-level mcpServers object exists.

    line 28
  4. Distinctive:

    Write JSON with 2-space indentation.

    line 30
  5. Distinctive:

    If writes fail on macOS, prompt for Full Disk Access and ensure the client app is closed before writing.

    line 49
  6. Distinctive:

    If Trae/Trae-CN installs fail due to permissions or other errors, generate a deep link and return it as a fallback.

    line 51
  7. Very specific:

    If trae or trae-cn cannot write to disk, generate a schema link for the user to open in Trae:

    line 55
  8. Very specific:

    Use references/clients.md for client IDs, aliases, and known config paths.

    line 67

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

MCP Installer

Overview

Install MCP server configurations across multiple clients and OSes with a CLI-first workflow and a safe manual fallback.

Workflow

1) Collect inputs

  • Identify the target client ID and OS; use references/clients.md for the supported list and known paths.
  • Gather MCP server definitions (name -> config object) and decide whether to override existing entries.
  • If any provided MCP config includes secrets (for example fields like key, token, apiKey, accessToken, secret, password, or Authorization, or env vars like *_KEY, *_TOKEN), pause the install and ask the user to obtain/provide those values first. Only proceed once the user confirms they have the required credentials (or requests installing a placeholder and will fill it in later).

2) Prefer @smithery/cli (when supported)

  • Run npx @smithery/cli --help and npx @smithery/cli install --help to confirm the exact syntax.
  • Use the CLI install flow for the client ID; if the CLI rejects the client or fails, fall back to manual edits.

3) Manual install (JSON config fallback)

  • Locate the config file path (see references/clients.md). If unknown, search standard config dirs for files like mcp.json, mcp_config.json, claude_desktop_config.json, or cline_mcp_settings.json.
  • Read existing JSON and ensure a top-level mcpServers object exists.
  • Merge servers: add new names; replace existing when override is true; preserve all other keys.
  • Write JSON with 2-space indentation.

Example shape:

{
  "mcpServers": {
    "example-server": {
      "command": "npx",
      "args": ["-y", "example-mcp@latest"],
      "env": {}
    }
  }
}

4) Verify and troubleshoot

  • Restart the client and confirm the MCP servers appear.
  • If writes fail on macOS, prompt for Full Disk Access and ensure the client app is closed before writing.
  • For CLI-only clients (for example claude-code), use the client CLI help to locate its MCP install/config command.
  • If Trae/Trae-CN installs fail due to permissions or other errors, generate a deep link and return it as a fallback.

5) Trae deep-link fallback

If trae or trae-cn cannot write to disk, generate a schema link for the user to open in Trae:

  • trae: trae://trae.ai-ide/mcp-import?type=${TYPE}&name=${NAME}&config=${BASE64_ENCODED_CONFIG}
  • trae-cn: trae-cn://trae.ai-ide/mcp-import?type=${TYPE}&name=${NAME}&config=${BASE64_ENCODED_CONFIG}

Rules:

  • TYPE is the MCP transport type (for example streamable-http).
  • NAME is the server name shown in Trae.
  • config is the base64-encoded JSON config object for that server (no whitespace, UTF-8 bytes).

References

  • Use references/clients.md for client IDs, aliases, and known config paths.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.