Four tools with basic descriptions and parameter schemas, but significant gaps in LLM-optimization and error handling. Tool names lack action verbs (refresh, sessions, session, status are weak). Descriptions are present but generic (10-50 chars for most). Parameter schemas exist but lack enums for constrained inputs (event, type, limit). No output schemas documented. No error recovery guidance. No tool annotations. Naming does not follow verb_noun pattern, 'refresh' and 'sessions' are ambiguous without context.

• Rename tools to follow verb_noun pattern: 'refresh' → 'check_catalog_updates', 'sessions' → 'search_sessions', 'session' → 'get_session_details', 'status' → 'get_event_status'.
• Add enum constraints for 'event' parameter across all tools: {"enum": ["build-2025", "ignite-2025", "build-2026"]} instead of free-form string.
• Document output schemas for each tool. E.g., sessions returns {"results": [{"sessionCode", "title", "speakers", "event"}], "total": number, "limit": number}.
• Add parameter constraints: 'limit' (min: 1, max: 100, default: 10), 'code' (pattern: '^[A-Z]{2,10}\d{1,5}$'), 'query' (min length: 1, description: 'Searches title, description, product, tags, topic, languages, speakers').
• Add error recovery guidance to 'sessions' description: 'Requires at least one of: --query, --tech, or --speaker. If search returns no results, try broader keywords or check available events with get_event_status.'