Server has 6 tools with complete input schemas and descriptions. Naming follows verb_noun convention (execute_query, get_rules, list_metrics). Descriptions are present but brief (avg 80 chars, below 194-char baseline). Parameters have types and descriptions. No output schemas documented. Error handling is minimal, make_prometheus_request() raises ValueError but provides no recovery guidance. Missing tool annotations (readOnlyHint, idempotentHint). All tools are read-only, which is safe, but this is not declared in tool metadata.

• Document output schemas for all tools. E.g., execute_query returns {resultType: string, result: array of {metric: object, value: [timestamp, value]}}. Include field types and meanings.
• Enhance error messages with recovery guidance. E.g., 'Invalid PromQL query: syntax error at position 42. Check query syntax at https://prometheus.io/docs/prometheus/latest/querying/basics/. Try: <corrected_query_example>'
• Expand tool descriptions to 150 - 250 chars. Add WHEN to use and dependencies. E.g., list_metrics(): 'Discover available metric names. Call this first to explore what data Prometheus has before writing queries. Returns all __name__ label values.'
• Add tool annotations: readOnlyHint=true and idempotentHint=true for all tools. This signals to LLMs that these calls are safe to retry and have no side effects.
• Add pagination to list_metrics() and get_labels(). Accept limit (default 100, max 1000) and offset parameters. Return total_count so LLM knows if more results exist.