Server has 8 well-named tools with consistent verb_noun patterns (get*, lookup*). All tools have descriptions (avg 95 chars, within 10-1024 baseline). Input schemas are present with type definitions and parameter descriptions. However, output schemas are not documented in the source code, LLMs cannot see what fields to expect from responses. Error handling is centralized via wrapToolErrorHandling() but lacks recovery guidance (e.g., 'try search_users first'). No tool annotations (readOnlyHint, destructiveHint) despite all tools being READ_ONLY. Parameter descriptions are adequate but could be more prescriptive about formats and constraints.

• Add output schema documentation to every tool. Document the structure of responses (e.g., 'Returns {portfolio: {accounts: [{id, positions: [{symbol, quantity, price}]}]}}') so LLMs know what fields to extract.
• Add tool annotations: mark all 8 tools with readOnlyHint=true in the tool definition. This signals to LLMs that these tools are safe to call without confirmation.
• Enhance error messages with recovery guidance. When auth fails, suggest 'Generate a new refresh token at https://www.questrade.com/api'. When a symbol is not found, suggest 'Try lookupSymbol() with a partial name.'
• Add format constraints to parameter descriptions. E.g., 'accountId: Questrade account number (8 digits, e.g., 12345678)' and 'ticker: Stock ticker symbol (1-5 uppercase letters, e.g., AAPL)'.
• Document pagination and result limits. If getPortfolio returns all accounts, state the limit or add offset/limit parameters. If results can exceed 100 items, add a 'limit' parameter (default 20, max 100).

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

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