33 tools with consistent verb-noun naming (get_, list_) and complete input schemas. All tools have descriptions (10-80 chars, mostly adequate). Parameters are typed and described. However, descriptions are uniformly terse (avg ~40 chars vs baseline 194 chars), lacking WHEN/WHY context and recovery guidance. No output schemas documented. No error handling guidance. All tools are READ_ONLY, reducing risk but also limiting composition patterns. Missing pagination guidance despite list_* tools. No tool annotations (readOnlyHint, etc.).

• Expand tool descriptions to 100-200 chars. Include WHEN to call (e.g., 'Call this after list_clusters to get detailed resource metrics'), WHAT it returns (e.g., 'Returns CPU, memory, storage, and IOPS utilization percentages'), and any prerequisites (e.g., 'Requires cluster ID from list_clusters').
• Document output schemas for all tools. Specify which fields are returned, their types, and which are safe to pass to downstream tools. E.g., list_clusters should document that it returns cluster_id, cluster_name, and datacenter_id.
• Add pagination guidance to list_* tools. Specify result limits (e.g., 'Returns up to 50 items'), mention offset/limit parameters if supported, and explain how to fetch additional results.
• Add error recovery hints to get_* and detail tools. E.g., 'If cluster not found, verify the cluster ID with list_clusters or check the vcenter_name parameter.'
• Declare tool annotations in schemas. Add readOnlyHint: true to all 33 tools to signal safety to LLMs and enable optimizations like caching and parallel execution.

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

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