Server has 4 tools with basic descriptions and partial schemas. Tool names follow verb_noun convention (analyze_, configure_, get_, data_broker_). However, descriptions lack specificity on WHEN to use each tool vs. alternatives, and parameter descriptions are minimal. Input schemas are present but incomplete: 'configurations' parameter uses generic List[Dict[str, Any]] instead of a structured schema with typed fields. No output schemas documented. Error handling exists but lacks recovery guidance. Missing tool annotations (readOnlyHint, destructiveHint). No evidence of idempotency guarantees or pagination support.

• Define 'configurations' parameter as a structured array of objects with explicit fields: expiration (string, pattern YYYYMMDD), strike_range (array of numbers, min 1, max 100), include_both_types (boolean, default true). Replace List[Dict[str, Any]] with proper JSON Schema.
• Document output schema for each tool. E.g., analyze_options_flow returns {monitored_contracts: [...], patterns: [...], institutional_bias: string, directional_signals: [...], most_active_strikes: [...]}. Include field types and descriptions.
• Add WHEN/WHY context to tool descriptions. E.g., 'Call this AFTER configuring monitoring with configure_options_monitoring_tool. Use to detect sweeps, blocks, and unusual volume patterns.' Clarify tool ordering and dependencies.
• Add tool annotations: mark configure_options_monitoring_tool with destructiveHint=true (modifies state), mark analyze_options_flow and get_monitoring_status_tool with readOnlyHint=true. Helps LLMs reason about side effects.
• Enhance error responses with actionable recovery. E.g., 'Data broker not running. Try: (1) call data_broker_health_check to verify connectivity, (2) check broker logs, (3) retry in 30 seconds.' Guide LLM's next step.

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

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