Three well-named, verb-prefixed tools with complete input schemas and clear descriptions. All parameters are typed and described. Tool names (list_wms_layers, get_wms_map, query_wfs_features) follow verb_noun convention and are unambiguous. Descriptions are substantive (100-200 chars), explaining what each tool does and what it returns. Input schemas use proper JSON Schema with types, descriptions, and sensible defaults. However, output schemas are not formally documented in the tool definitions, LayerInfo is a Pydantic model but not exposed in the schema metadata. Error handling is present (wrap_connection_errors context manager) but not integrated into tool responses; agents see raw exceptions rather than actionable recovery guidance. No tool annotations (readOnlyHint, destructiveHint, idempotentHint) despite all three being READ_ONLY operations.

• Add tool annotations: @mcp.tool(readOnly=True) to all three tools to declare their read-only nature and enable client optimizations.
• Document output schemas explicitly. For list_wms_layers, add a return type hint and docstring describing LayerInfo fields. For query_wfs_features, document the GeoJSON FeatureCollection structure (type, features array, bbox, etc.).
• Wrap tool implementations with try-except to catch OGCError and return structured error responses. E.g., 'Layer "xyz" not found. Available layers: layer1, layer2, layer3' instead of raising an exception.
• For get_wms_map, document that the return value is binary image data and specify supported formats (png, jpeg, gif, etc.) in the description.
• Add pagination guidance to query_wfs_features: document that max_features defaults to 100 and explain how to retrieve additional features if the result is truncated.