pomelli-mcp exposes 12 browser automation tools with minimal documentation. Tool names are action-oriented (pomelli_click, pomelli_fill, pomelli_goto) but descriptions are inconsistent in depth. Most tools lack parameter descriptions entirely, e.g., pomelli_fill has 'value' and 'selector' params with descriptions, but pomelli_click's 'text' param has only a bare description. No output schemas are documented. Error handling is absent; tools will fail silently or return raw exceptions. The server is STDIO-only, capping protocol readiness at 50. Overall definition quality is below median for production tools.

• Document output schemas for all tools. E.g., pomelli_status should return {logged_in: bool, geo_available: bool, ready: bool, screenshot_path: string}. pomelli_snapshot should return {elements: [{type, text, selector}], screenshot_path: string}.
• Expand parameter descriptions with format, constraints, and examples. E.g., pomelli_click: 'text: Case-insensitive substring match against visible text or aria-label. E.g., "Generate Brief" matches "generate brief" button.'
• Add error handling with recovery guidance. E.g., 'Element not found. Try pomelli_snapshot() to see current UI, then retry with correct text.'
• Document tool sequencing and prerequisites. E.g., pomelli_download_assets: 'Call pomelli_wait_done() first to ensure generation is complete. Scroll via pomelli_goto() or pomelli_press() before calling this.'
• Constrain pomelli_js with a whitelist of safe operations or remove it in favor of explicit tools (e.g., pomelli_get_element_count, pomelli_evaluate_condition).