Two tools with basic schemas and descriptions, but significant gaps in parameter documentation and error handling. get_task has optional parameters with minimal guidance on defaults and dependencies. check_pricing lacks clarity on what 'model' and 'action' mean in context. Neither tool documents output structure or error recovery paths. Descriptions are present but generic (avg ~80 chars, below the 194-char baseline for A+ tools). No parameter descriptions visible in the schema, only inline type hints. No error handling guidance or recovery suggestions.

• Add explicit parameter descriptions to both tools. For get_task: 'task_id (string, required): The task ID returned by a create task call. action (string, optional): The endpoint name where the task was created. If omitted and only one async endpoint exists, defaults to that endpoint; if multiple exist, this parameter is required.'
• Document the response schema for each tool. E.g., get_task returns {task_id, status, task} where task contains {id, status, result, created_at, ...}. check_pricing returns {model, pricing_url, error?}.
• Add error handling guidance. E.g., 'If task not found, the task may have expired or the task_id may be invalid. Verify the task_id and try again. If the error persists, the task may no longer be available.'
• Expand tool descriptions to 100-150 chars explaining the workflow. E.g., 'get_task: Poll the status of an async image generation task. Call this after creating a task to check if it is complete and retrieve the result URLs.'
• Declare valid enum values for 'action' parameter. Instead of a free-form string, use z.enum([...endpoints]) to make valid options discoverable and prevent hallucinated values.