Three well-named, read-only tools with clear descriptions and complete input schemas. All tools follow verb_noun naming (search_, query_, layer_). Descriptions are substantive (150 - 250 chars) and explain purpose, prerequisites, and return values. Input schemas include type definitions and parameter descriptions. However, output schemas are not documented, callers cannot predict response structure. Error handling is minimal (generic HTTP status errors). No pagination guidance for search_datasets despite potentially large result sets. Missing required field declarations in some schemas.

• Document output schemas for all three tools. Example: search_datasets returns {total: number, count: number, datasets: [{id, name, summary, org, org_id, record_count, type, url}]}. This enables LLMs to chain results into query_layer calls.
• Add pagination guidance to search_datasets: 'Returns up to 50 datasets per page. Use limit to control page size (1 - 50, default 20). Check total field to determine if more results exist; implement client-side pagination by adjusting limit/offset in subsequent calls.'
• Replace example SQL values in query_layer description with formal constraints. Instead of 'e.g. "STATE = \'CA\'..."', state: 'SQL WHERE clause following ArcGIS REST API syntax. Common patterns: field_name = value, field_name >= value, field_name IN (val1, val2).'
• Enhance error messages with recovery guidance. E.g., 'ArcGIS Hub returned 404. Verify org_id is correct (default: Ne8d9gKn5SJ3eAaw for Glasgow City Council). Try search_datasets with a broader keyword.' This guides LLM retry logic.
• Add 'required' array to search_datasets schema: required: ['query']. Clarify that org_id and limit are optional with sensible defaults.