Server has 21 tools with basic verb-noun naming (get_, list_, send_, etc.), but critical gaps undermine quality. Most tool descriptions are present but generic (10-50 chars), falling below the 50-200 char LLM-optimized baseline. Input schemas are visible for most tools, but many lack parameter descriptions entirely. No output schemas documented. Error handling is minimal, no recovery guidance, no categorization of retryable vs fatal errors. No tool annotations (readOnlyHint, destructiveHint, idempotentHint) despite clear risk levels. Parameter validation rules not documented in descriptions. No pagination support despite tools like list_players, list_backups, get_server_logs that could return large result sets.

• Expand tool descriptions to 50-200 characters. Include WHEN to use the tool and WHAT it returns. E.g., 'Get current server status (online/offline, player count, uptime). Call this first to check if the server is running before sending commands.'
• Document output schemas for all tools. Specify field names, types, and whether they are required. E.g., list_players returns [{player_id: string, player_name: string, join_time: ISO8601}].
• Add tool annotations: mark read-only tools with readOnlyHint=true, destructive tools (ban_player, restore_backup) with destructiveHint=true, and idempotent tools with idempotentHint=true.
• Add pagination to list tools: accept limit (1-100, default 20) and offset/cursor parameters. Return total_count and next_cursor so agents can fetch large result sets incrementally.
• Add error recovery guidance: document which errors are retryable (network timeouts, 503 service unavailable) vs user-fixable (invalid player_name, backup_id not found). Provide actionable next steps in error messages.

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

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