Run Diagnostics: AI Health Check for Your Mac

run_diagnostics runs a fast, read-only health check of every LMCP integration on your Mac and reports back in plain language: what's connected, what's broken, and how to fix it. Instead of digging through System Settings, permissions panels, and config files, you just ask your AI assistant. A typical prompt looks like: "Run diagnostics on my Mac and tell me which apps aren't connected yet." The AI checks Mail, Calendar, Contacts, Reminders, Messages, Notes, Safari, Teams, Slack, WhatsApp and the rest of the 150+ tools, then hands you a clear status list with next steps.

This is the first thing to run when something isn't working. If a calendar event won't create or email search comes back empty, run_diagnostics tells you whether it's a missing macOS permission (TCC), an app that isn't signed in, or an integration you simply haven't connected yet — and what to click to resolve it. Download LMCP to get started.

Which AI agents work?

run_diagnostics works across every major AI client. Desktop apps auto-configure over local stdio the moment LMCP is installed:

• Claude Desktop — auto-configured via local stdio
• Cursor — auto-configured via local stdio
• VS Code (GitHub Copilot) — auto-configured via local stdio
• Windsurf — auto-configured via local stdio
• Zed — auto-configured via local stdio

Web-based AIs connect through the LMCP Cloud Relay connector, which securely bridges the chat to the server running on your Mac:

• ChatGPT — via LMCP Cloud Relay connector
• Claude.ai (web) — via LMCP Cloud Relay connector
• Grok — via LMCP Cloud Relay connector
• Perplexity — via LMCP Cloud Relay connector

Automation

run_diagnostics is the natural opener for any multi-step workflow. Chain it ahead of real work so the AI knows what it can rely on: run diagnostics, confirm Calendar and Contacts are healthy, then schedule a meeting and email the attendees. If diagnostics flags Mail as disconnected, the AI can skip the email step and tell you why instead of failing silently halfway through. It pairs especially well with get_config and lmcp_state for a complete picture before bulk operations.

Context

The check runs against your real Mac data and your real apps — the actual Mail accounts you're signed into, the calendars you use, the contacts in your address book. There's no sandbox or demo data. That means the diagnosis reflects exactly what your AI assistant will see when it acts, so a green result genuinely means the next prompt will work.

Productivity

Permission and connection issues are the number-one reason an AI task stalls on a Mac. Manually tracing one usually means opening System Settings, hunting through Privacy & Security, checking which account an app is signed into, and re-reading docs. run_diagnostics collapses all of that into a single sentence and a few seconds — turning a 15-minute troubleshooting detour into one question. Run it once after install and you're confident every integration is ready.

Privacy & GDPR

Everything in run_diagnostics happens locally on your Mac. The health check reads your app and permission status on-device and returns it straight to your AI client — nothing is uploaded, copied to a server, or stored in the cloud. There are no API keys to hand over and no third-party data processor in the loop. This is privacy by architecture, not policy. Learn more about how LMCP is GDPR-compliant by architecture.