Three tools with complete input schemas and descriptions, but significant gaps in output documentation and error handling. search_patterns and upload_chart have good parameter constraints (enums, descriptions), but lack documented return schemas. list_patterns is trivial (no params). No tool annotations (readOnlyHint, destructiveHint). Error handling exists in code but not surfaced in tool definitions. Descriptions are adequate (100-150 chars) but lack WHEN/WHY guidance. No pagination or result limits documented despite search_patterns returning 'top_n' results.

• Document output schema for search_patterns: 'Returns {items: [{symbol, score, pattern_name, timeframe, ...}], total_count, next_cursor}' so agents know what fields to extract.
• Add bounds to top_n parameter: 'Number of top results to return (1 - 10, default 5)' instead of just 'default 5, max 10'.
• Add readOnlyHint: true to all three tool definitions since they are read-only queries with no side effects.
• Expand list_patterns description: 'Returns all ~50 supported pattern templates with names, aliases, and descriptions. Call this first to understand available patterns before searching.'
• Document error cases in upload_chart: 'If image_path does not exist or is not a valid image, returns error with the invalid path. Ensure file is PNG/JPG under 2MB.'

Inferred effective spec: 2025-06-18+.

• Relies on Stateful initialize / Mcp-Session-Id (removed; protocol is stateless) - make each request self-contained