Loading ToolBench
An MCP server that provides tools to search, discover, and manage registered MCP servers and their capabilities in a centralized catalog
Five tools with complete input schemas and descriptions. Naming follows verb_noun pattern (catalog_search, catalog_get_server, etc.). Descriptions are substantive (150-250 chars), explaining WHAT, WHEN, and WHY. All parameters have types and descriptions. However, output schemas are not documented in the provided source, only input schemas are visible. No error handling guidance or recovery patterns evident. No tool annotations (readOnlyHint). All tools are READ_ONLY, which is appropriate, but this should be declared via annotations for protocol compliance.
Check the health status of a registered MCP server. Use this tool to verify if a specific MCP server is currently operational and responding. This will return the current status along with recent health history. Useful for: - Verifying a server is online before using its tools - Monitoring server availability - Troubleshooting connectivity issues
Get detailed information about a specific MCP server. Use this tool to retrieve comprehensive details about a registered MCP server, including its capabilities, tools, resources, and health status. You can specify what additional information to include.
Get detailed information about a specific tool from the catalog. Use this tool to retrieve comprehensive details about a tool's input/output schemas, description, and which MCP server provides it. This is useful when you need to understand how to use a specific tool or want to see its exact input requirements and expected outputs.
List all available capabilities in the MCP catalog. Use this tool to browse what types of tools and services are available across all registered MCP servers. Capabilities are grouped by category (database, filesystem, search, CRM, etc.). This is useful when you want to explore what's available without a specific search query.
Output schemas not documented. Tool descriptions explain inputs but not what fields/structure the response contains. LLMs cannot plan downstream calls or extract data without knowing response shape.
No tool annotations (readOnlyHint, destructiveHint, idempotentHint). All five tools are READ_ONLY and idempotent, but this is not declared in the tool definition. Agents cannot infer safety properties.
No error handling guidance. Descriptions do not explain what errors can occur or how to recover. E.g., catalog_get_server does not document what happens if serverId is invalid or not found.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | B | 71 | 2025-06-18+ | v2 |
Search the MCP catalog for servers and capabilities. Use this tool to find MCP servers by capability, category, or keywords. Perfect for discovering tools that can perform specific tasks like: - "I need a tool that can query Salesforce" - "Find database query tools" - "Search for file system capabilities" The search covers server names, descriptions, capability names, and tags.
Parameter 'q' in catalog_search is required but description does not specify minimum length, character restrictions, or examples of valid queries. Agents may pass empty or malformed strings.
catalog_list_capabilities and catalog_search both offer pagination (limit, offset) but do not document total count or next_cursor in response. Agents cannot determine if more results exist.