Three well-named, read-only tools with clear descriptions and complete input schemas. All tools follow verb_noun naming (search_, query_, layer_info). Descriptions are concise (100 - 200 chars) and explain WHAT, WHEN, and WHAT-TO-DO-NEXT. Input schemas include types and descriptions for all parameters. However, output schemas are not documented, LLMs cannot predict response structure for downstream chaining. Parameter constraints (e.g., limit 1 - 50, 1 - 2000) are documented in descriptions but not enforced as JSON Schema constraints. No error handling guidance or recovery paths. Security is solid (SSRF guard in layerUrl, no secrets in params), but no permission gates or audit logging.

• Document output schemas for all three tools. Example for search_datasets: 'Returns {total: number, count: number, datasets: [{id, name, summary, org, org_id, record_count, type, url}]}'. This enables LLMs to chain tools and extract IDs for downstream calls.
• Add error handling with recovery hints. Replace 'ArcGIS Hub: 404' with 'Dataset not found. Try search_datasets with a different keyword or check the org_id.' Categorize errors as retryable (network timeout) vs user-fixable (invalid URL format).
• Add JSON Schema constraints for numeric parameters: limit with minimum/maximum, offset with minimum. Example: 'limit: {type: number, minimum: 1, maximum: 50, default: 20}'.
• Strip verbose ArcGIS metadata from query_layer response. Return only features array with attributes and geometry; omit exceededTransferLimit, globalIdFieldName, etc. Cap results at 100 features and document this limit in the description.
• Add next_cursor or offset guidance to search_datasets response when total > count. Example: 'If total > count, call search_datasets again with limit and org_id to fetch the next page.' This enables pagination without forcing LLM guesswork.