AI Tools Setup Assistant
Set up, connect, validate, and repair Claude Desktop and MCP servers using Desktop Commander.
The skill explicitly requires backing up configs, verifying changes, not leaking secrets, and only adding trusted MCP servers, showing good security practices. However, it lacks user confirmation mechanisms, data-flow transparency, and clear external impact statements; permissions are not fully controlled, so deductions apply.
The skill provides clear diagnostic steps and troubleshooting checklists, but it relies on external docs and network resources, has no automated tests covering key paths, and static review cannot verify actual reliability, so deductions apply.
The skill clearly defines target scenarios (configuring Claude Desktop and MCP servers) and trigger symptoms, provides multi-platform paths, but limited capability boundaries and non-fit ranges, and relies on overseas services (e.g., Nous Portal, OpenClaw) which may be inaccessible to mainland users, so deductions apply.
Documentation is well-structured with version, audience, description, and detailed reference files, but lacks changelog, maintenance responsibility, and known limitations; versioning may not align with content, so deductions apply.
The skill's logic is complete and steps executable, but static review cannot verify actual effect, external dependencies may change, and output is unverified, so deductions apply.
The skill references external docs and GitHub but no concrete test results or third-party verification evidence; static review cannot independently reproduce, so deductions apply.
- The skill relies on MCP servers and external services (e.g., Nous Portal, OpenClaw) that may be inaccessible from mainland China networks; consider network reachability.
- Configuration modifications mentioned may affect system environment; ensure user confirmation and backup to avoid mistakes.
- The skill lacks explicit error handling and rollback mechanisms; exercise caution when encountering anomalies.
What does this skill do, and when should you use it?
This skill guides an AI agent to help users install, configure, and repair AI tooling on their own machine, focusing on configuration and troubleshooting rather than clean installs. It covers installing Claude Desktop, adding and fixing MCP servers, diagnosing connection issues, and wiring local AI gateways like OpenClaw and Hermes. The skill emphasizes safety practices such as backing up configs, verifying changes, and keeping secrets secure.
The skill reads config files like mcp.json and claude_desktop_config.json to inspect current configuration. It uses terminal commands (via start_process) to check system info, validate API keys, test model endpoints, and confirm services are running by listing processes and ports. It edits config files (using edit_block or write_file) to add or modify MCP server entries, backs up files before changes, and validates JSON syntax. It also provides structured troubleshooting steps for diagnosing 'agent not responding' issues, such as checking authentication, model support, process status, and logs.
- User reports 'Claude can't see my tools' - Skill checks MCP config and verifies servers are correctly listed.
- User wants to 'add an MCP server' - Skill edits the appropriate config file and verifies the connection.
- User says 'my MCP server isn't showing up' - Skill diagnoses config errors and process issues.
- User wants to 'wire up Ollama or LM Studio' - Skill assists with configuring local model providers.
- User needs to 'restart Claude Desktop' - Skill guides safe restart via process management.
- User is having issues with OpenClaw or Hermes gateways - Skill provides configuration and diagnostic steps.
What are this skill's strengths and limitations?
- Safety-focused: backs up configs, avoids echoing secrets, verifies changes.
- Clear troubleshooting workflow covering common symptoms.
- Covers multiple tools (Claude Desktop, MCP, OpenClaw, Hermes).
- Repeatable results with suggestions to save steps.
- Installation steps for the skill itself are undocumented; requires manual setup.
- Depends on an external MCP server (Desktop Commander), adding installation overhead.
- No explicit test suite or version history.
- Cross-platform behavior may vary by OS; not verified.
How do you install this skill?
This skill resides in the Desktop Commander MCP repository. It is part of a broader collection that works with MCP clients that support Agent Skills (like Cursor). The installation steps for the skill itself are not documented in the provided source material. However, the repo provides standard MCP installation methods (e.g., via npx or manual config). To add this skill to your agent, you may need to copy its SKILL.md file into your skills directory or use a client that supports skill loading.
How do you use this skill?
Once triggered, the skill will read the relevant reference file (e.g., references/claude-desktop-mcp.md) for exact paths and commands. Follow the general workflow: 1. Clarify the goal and platform; 2. Inspect current state (read configs, list processes); 3. Make the smallest change, backing up first; 4. Verify the result; 5. Summarize what changed and how to undo. Example trigger prompts: 'Fix my MCP config' or 'Claude can't see my tools'.