This is a configuration file that proxies to 6 external MCP servers via HTTP. No tool definitions, schemas, or descriptions are visible in the source code provided. The .mcp.json file only contains server URLs and headers, it does not define tools, parameters, or output schemas. Tool names and descriptions are inferred from the configuration keys and brief labels, not from explicit tool registration. Per the hard scoring rules, inferred tool definitions cap at 50 per tool; since NO actual schemas or parameter definitions are visible anywhere in the provided source, all tools score 0 on schema. Descriptions are generic labels (e.g. 'AWS Knowledge MCP tool, provides access to AWS documentation') under 50 characters, scoring 0-20. This server is a thin proxy wrapper with no intrinsic tool quality, it delegates entirely to upstream services.

• Define explicit tool schemas in the MCP server implementation (not just config). Each tool must have a name, description (50 - 200 chars), input schema with typed parameters, and documented output schema.
• Rename tools with action verbs: search_aws_documentation, list_microsoft_learn_resources, search_google_developer_docs, query_datacommons, search_government_repositories, list_uk_tenders. Remove wildcard suffixes.
• Expand descriptions to 50 - 200 characters explaining WHAT the tool does, WHEN to use it (e.g. 'Use this to find AWS service documentation and best practices'), and what it returns (e.g. 'Returns matching docs with title, URL, and summary').
• Document output schemas for each tool. Specify fields (title, url, summary, relevance_score), types (string, number, array), and whether results are paginated.
• Add parameter descriptions for all inputs (search_query, limit, offset, filter_by_service, etc.). Include constraints: 'search_query: 1 - 500 chars, required'; 'limit: 1 - 100, default 20'.