Submit a QA Report to LMCP from AI on Mac

The submit_qa_report tool lets an AI assistant send an automated QA run report straight to the LMCP team. It's the reporting endpoint that Cowork and other agents use for continuous quality monitoring of Local MCP — capturing which tools were exercised, what passed, what failed, and the surrounding context, then submitting it in one clean call. A typical prompt looks like: "Run a read-only sweep across Mail, Calendar, Contacts and Notes, then submit a QA report with the results." Your AI does the sweep against your real Mac apps and files the report for you.

Because LMCP is a free, native macOS MCP server, this all happens on your machine. The AI talks to your local apps directly and the report is submitted on your behalf. Download LMCP to get started.

Which AI agents work?

LMCP exposes submit_qa_report alongside 150+ tools to every major MCP-capable assistant:

• Desktop clients — Claude Desktop, Cursor, VS Code (GitHub Copilot), Windsurf and Zed auto-configure through a local stdio connection. No keys, no setup beyond install.
• Web AIs — ChatGPT, Claude.ai (web), Grok and Perplexity connect through the LMCP Cloud Relay connector, which securely bridges the web assistant to the server running on your Mac.

Automation

This tool is designed to chain. An agent first runs diagnostic and read tools — run_diagnostics, lmcp_state, get_audit_log, plus live read-only calls across Mail, Calendar, Contacts, Notes, Reminders, Messages and the rest — then bundles the outcomes and calls submit_qa_report to deliver the verdict. It turns a multi-step verification pass into a single submitted artifact, so quality signals reach the LMCP team automatically instead of living in a chat log.

Context

The value of a QA report is that it reflects your real environment. The AI works against the actual apps and data on your Mac — your mail accounts, your calendars, your local files — so the report captures how LMCP behaves on a live machine, not a synthetic fixture. That real-world context is exactly what makes continuous monitoring useful: it surfaces issues that only appear with genuine accounts and permissions.

Productivity

Filing a quality report by hand means running each check, copying outputs, formatting findings and sending them somewhere. With submit_qa_report, you say one sentence and the agent runs the sweep and submits the structured result in seconds. For continuous monitoring this compounds: a scheduled agent can validate the whole tool surface on a cadence and report regressions without anyone watching, freeing you to act on findings instead of gathering them.

Privacy & GDPR

Everything runs locally on your Mac. The AI reads your apps through the operating system's own permission system, and no copy of your mail, calendar, contacts or files is stored on any server. Only the QA report itself — a summary of which tool checks passed or failed — is submitted to the LMCP team; your underlying data never leaves the machine. There are no API keys to hand over and nothing sits in the cloud. See GDPR-compliant by architecture for how this local-first design keeps you compliant by default.