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. Error handling is minimal (generic HTTP status errors). No parameter constraints (enums, ranges) despite opportunities (e.g., limit 1 - 50 for search_datasets). SSRF guard in layerUrl() is a security plus. Missing: per-parameter validation guidance, output schema documentation, error recovery hints.

• 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 extract and chain results.
• Add enum constraints to parameters. E.g., limit: {type: 'number', minimum: 1, maximum: 50, default: 20}. For query_layer, add pattern validation to 'where' (basic SQL syntax check) and 'out_fields' (comma-separated field names or '*').
• Replace generic error messages with actionable guidance. E.g., 'ArcGIS Hub returned 404: dataset not found. Try search_datasets with a different keyword.' or 'ArcGIS query failed (500): service may be temporarily unavailable. Retry in 30 seconds.'
• Add timeout and rate-limit handling. Set explicit fetch() timeouts (e.g., 10s) and return a clear error: 'Request timed out after 10s. The ArcGIS service may be slow; try again in a moment.' Consider per-minute call limits to prevent agent runaway.
• Clarify chaining in search_datasets description: 'Each dataset includes a Feature Service url, pass it directly to query_layer() or layer_info() to query or inspect the schema.' Explicitly state that url is always present.