Loading ToolBench
MCP server for browser automation and control using AI-powered browser (FBrowser) with CEF3 backend
This MCP server has severe definition quality issues across nearly all dimensions. Tool names are inconsistent and lack action verbs (e.g., 'ConsoleOutput' vs verb_noun convention). Descriptions are minimal (10-50 chars) and lack context about WHEN to use each tool or what it returns. Parameter schemas are present but descriptions are sparse and many lack actionable guidance. No documented output schemas. Error handling is not evident. The 'rg_KaiFaZheXiaoXi_VIP_*' tools use Chinese pinyin identifiers, making them opaque to English-speaking LLMs and violating naming conventions. Most critically, 5 tools (OnDevToolsMessage, OnDevToolsMethodResult, OnDevToolsEvent, OnDevToolsAgentAttached, OnDevToolsAgentDetached) accept 'object' type parameters without schema definitions, which violates JSON Schema best practices and prevents LLM parameter validation.
Check if a directory exists
Output text to console
Convert integer value to TLS version
Handle DevTools agent attachment event
Handle DevTools agent detachment event
Handle receipt of a DevTools protocol event
Handle receipt of a DevTools protocol message
Handle attempted execution result of a DevTools protocol method
CRITICAL: 5 tools (OnDevToolsMessage, OnDevToolsMethodResult, OnDevToolsEvent, OnDevToolsAgentAttached, OnDevToolsAgentDetached) accept 'object' type parameters with no schema definition. LLMs cannot validate or construct these parameters correctly.
CRITICAL: 5 tools use Chinese pinyin names (rg_KaiFaZheXiaoXi_VIP_*) that are opaque to English-speaking LLMs and violate verb_noun naming convention. These should be renamed to English action verbs (e.g., 'handle_developer_message_received').
HIGH: No tool descriptions explain WHAT the tool does, WHEN to call it, or WHAT it returns. All descriptions are 20-50 chars and lack actionable context. E.g., 'Output text to console' doesn't explain whether this is for logging, debugging, or user interaction.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 30 | 2026-07-28+ | v2 |
Developer message VIP receive event handler
Developer message VIP receive message handler
Developer message VIP detached handler
Developer message VIP attached handler
Developer message VIP execution completion handler
HIGH: Tool names do not follow verb_noun convention. 'ConsoleOutput' should be 'output_to_console', 'CheckDirIsExists' should be 'check_directory_exists'. This makes LLM intent inference difficult.
HIGH: No output schemas documented for any tool. LLMs cannot plan downstream tool calls or extract required fields (e.g., what does OnDevToolsMessage return?). This violates the documented output schema requirement.
MEDIUM: Parameter descriptions are minimal or missing. E.g., OnDevToolsMethodResult has 'Browser instance that originated the method call' but doesn't explain the type or constraints. Parameter descriptions under 30 chars provide insufficient context.
MEDIUM: No error handling documentation. Tools do not document what errors they return, when to retry, or how to recover. E.g., CheckDirIsExists doesn't explain what happens if the path is invalid or unreadable.
MEDIUM: 'object' type parameters lack nested schema definitions. E.g., OnDevToolsMessage accepts a 'browser' object but doesn't document what fields it contains or their types. This makes parameter construction impossible for LLMs.