Server has 5 tools with basic schemas and descriptions, but critical gaps prevent production use. All tools lack output schema documentation, LLMs cannot predict return structure. Descriptions are present but generic (avg 80 chars, below 194-char baseline). Parameters lack type annotations in schema (dict/list without JSON Schema structure). No error handling guidance, no parameter validation rules, no recovery paths. Tool names are verb-noun but descriptions don't explain WHEN to use each tool or dependencies between them. All tools are read-only (good), but no tool annotations (readOnlyHint) to signal this to clients. Return values are coerced to strings, losing structure.

• Add JSON Schema definitions for all input parameters. Replace dict/list with explicit properties, required arrays, and constraints (e.g., filter: {type: 'object', description: 'MongoDB filter document (e.g., {"status": "active"})', additionalProperties: true}).
• Document output schemas for each tool. E.g., query returns {type: 'array', items: {type: 'object'}, description: 'Array of matching documents'}. Specify field names users will see.
• Add readOnlyHint: true to all 5 tools via @mcp.tool(readOnlyHint=True) or equivalent FastMCP annotation.
• Expand descriptions to 100-200 chars. Include WHEN to use: 'Use query for simple filters; use aggregate for multi-stage transformations (group, sort, lookup).' Add format hints: 'filter accepts MongoDB query syntax, e.g., {"age": {"$gt": 18}}'.
• Implement error handling with recovery guidance. Catch pymongo.errors.* and return structured errors: {error: 'Collection not found', suggestion: 'Call listCollections() to see available collections', retryable: false}.

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

• Relies on Logging (deprecated) - log to stderr or use OpenTelemetry