Single tool 'collect_earnings_data' has a detailed description (195 chars) and clear input schema with two parameters (symbol: string, full: boolean). However, parameter descriptions lack actionable constraints (no format hints for symbol, no explanation of what 'full' includes beyond vague reference). Output schema is completely undocumented, no indication of what fields the response contains, their types, or how to chain this tool with others. No error handling guidance. Tool name is verb-noun compliant but lacks clarity on whether it performs a read-only fetch or triggers side effects. Risk marked READ_ONLY is good, but no recovery guidance if a symbol is invalid or data unavailable.

• Document the complete output schema with field names, types, and descriptions. Example: '{snapshot: {price: number, change: number}, income_statement: [{quarter: string, revenue: number, ...}], ...}'. Specify which fields are always present vs. conditional on 'full=true'.
• Add format constraint to 'symbol' parameter: 'Stock symbol in format TICKER or TICKER.EXCHANGE (e.g., AAPL, 09988.HK). 1-10 characters, alphanumeric + dot.'
• Expand 'full' parameter description: 'When false (default), returns core earnings data (snapshot, income statement, consensus, forecast, quote, index, rating, segments, news, kline). When true, adds balance sheet, cash flow, filings, industry valuation, peer comparison, and rating history. Note: full=true may take 2-5 seconds longer.'
• Add error handling section to description: 'Returns error if symbol not found or data unavailable. If symbol is invalid, try search_symbols() first. If rate-limited, retry after 60 seconds.'
• Clarify in description that this is a read-only parallel fetch with no side effects, and specify typical response time and data freshness (e.g., 'Real-time snapshot, income statement updated quarterly').