Critical failures across naming, descriptions, and schemas. 15 of 23 tools (65%) have empty input schemas with no visible parameter definitions. 8 tools have descriptions under 20 characters or are completely missing descriptions. Tool names lack clarity (e.g., 'chrome_keyboard', 'chrome_history', 'chrome_console' are vague about what they do). No output schemas documented. Error handling is absent, no recovery guidance or actionable error messages visible in code. This server is not production-ready.

• Add complete input schemas to all 8 empty-schema tools. For chrome_keyboard, define parameters like 'keys' (string array), 'delay' (number ms). For chrome_history, add 'query' (string), 'maxResults' (number 1-100). For chrome_console, add 'command' (string) and 'returnValue' (boolean).
• Write descriptions for all 8 undescribed tools. Example: 'chrome_keyboard: Send keyboard input (text, key presses, shortcuts) to the active browser tab. Use after clicking an input field.' Minimum 20 chars, maximum 200 chars per pattern:tool-description.
• Document output schemas for all 23 tools. Example for get_windows_and_tabs: {windows: [{id: number, tabs: [{id: number, url: string, title: string, active: boolean}]}]}. This enables LLM planning.
• Split 'chrome_fill_or_select' into two tools: chrome_fill_input (selector, value) and chrome_select_option (selector, value). Split 'chrome_go_back_or_forward' into chrome_go_back() and chrome_go_forward(). Each tool = one action.
• Add error handling and recovery guidance. For destructive tools (chrome_close_tabs, chrome_bookmark_delete), return actionable errors: 'Tab not found. Available tabs: [list]. Did you mean one of these?' For chrome_inject_script, validate script syntax before execution and return parse errors with line numbers.

Inferred effective spec: <=2025-11-25.

• Relies on Stateful initialize / Mcp-Session-Id (removed; protocol is stateless) - make each request self-contained