MCP Server for Flutter Tools - provides tools for monitoring, debugging, and controlling Flutter applications
Flutter Tools MCP exhibits significant definition quality gaps. All 9 tools have basic descriptions (10-50 chars), but most fall below the 194-char baseline for production tools. Input schemas are present but minimal: most parameters lack detailed descriptions, and several tools have empty input objects. No output schemas are documented. Error handling is absent, no recovery guidance, categorization, or actionable error messages. Tool naming follows verb_noun convention (list-, connect-, get-, take-, analyze-) but descriptions are too terse to guide LLM selection. No parameter constraints (enums, ranges, formats) beyond the single enum in get-performance-metrics. The server appears to be a thin wrapper around Flutter VM Service APIs without agent-oriented design.
Run a performance analysis on the Flutter app
Connect to a Flutter app by VM service URL
Connect to a specific Flutter app by ID
Retrieve logs from a connected Flutter app
Fetch network request data from a Flutter app
Get performance metrics from a Flutter app
Retrieve the widget tree structure of a Flutter app
Descriptions are too terse (10-50 chars vs. 194-char baseline). LLMs cannot determine when to select each tool or how it differs from similar tools. E.g., 'List all running Flutter applications' lacks context on what 'running' means, whether it includes backgrounded apps, or what fields are returned.
No output schemas documented. LLMs cannot plan downstream tool calls or extract required fields. E.g., does list-flutter-apps return app IDs, names, and connection status? What fields does get-app-logs return, raw text, structured entries with timestamps, or both?
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 41 | 2026-07-28+ | v2 |
List all running Flutter applications
Capture a screenshot of the Flutter app UI
Parameter descriptions are missing or minimal. 'appId' is described as 'The ID of the Flutter app to connect to' but lacks guidance on format, where to find it, or whether it matches the ID returned by list-flutter-apps. 'lines' and 'count' parameters have no range constraints (e.g., 1 - 1000).
No error handling or recovery guidance. If connect-to-app fails (app not found, connection refused, auth token invalid), the tool provides no actionable next step. LLMs cannot self-correct or suggest alternatives.
Tool composition unclear. connect-to-app and connect-by-url both establish connections but differ in input (ID vs. URL). Descriptions do not explain when to use each or whether they are mutually exclusive. No guidance on connection state management or whether multiple connections can coexist.