mcp-todo has 6 tools with complete input schemas and basic descriptions. However, descriptions are minimal (10-50 chars), parameter descriptions lack detail about constraints and formats, and output schemas are not documented. Tools follow verb_noun naming but lack guidance on when to use each tool vs. alternatives. No error recovery hints, no pagination support, and responses embed full state (todos array) on every call rather than structured, minimal output. Schemas are present but descriptions fall short of LLM-optimization standards (target 50-200 chars per pattern:tool-description).

• Expand tool descriptions to 50-200 characters. Example: 'add_todos: Add one or more todos to the list. Use clear_existing=true to reset the list before adding. Returns the newly created todos with their IDs.' This explains WHAT, WHEN, and the return value.
• Add parameter descriptions with format and constraint hints. E.g., 'texts: Array of todo text strings (1-500 chars each). Each text becomes a new todo item.' and 'ids: Array of todo ID strings (returned by add_todos or list_todos).'
• Document output schemas in tool definitions. Add a 'outputSchema' field describing the response structure: {success: boolean, added?: array of {id, text, done, inProgress}, todos: array of all current todos, ...}.
• Add error recovery hints to parameter descriptions. E.g., 'ids: Array of todo IDs. If an ID is not found, the response will indicate {found: false} for that ID, call list_todos to see available IDs.'
• Mark destructive operations explicitly. Update clear_todos description to: 'Clear all todos at once. WARNING: This is irreversible. Returns the count of todos deleted.' Consider adding a confirmation pattern (dry-run or explicit confirmation parameter) per pattern:confirmation-request.