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 194-char baseline). No tool annotations, no idempotency hints, no permission scopes declared. Error handling visible in code (friendlyError calls) but not surfaced in tool definitions.

• Add enum constraints to 'action' parameters. In get_task, if asynchronousEndpoints is known at registration time, declare action as z.enum([...endpoints]) instead of a free-form string. This self-documents valid values.
• Document output schemas in tool descriptions. E.g., 'Returns {task_id: string, status: string, task: object}' or link to a schema definition. This enables LLMs to extract and chain results.
• Expand check_pricing description to 100-150 chars: 'Look up the cost per token or per request for a specific Hailuo model on RunAPI. Use this before creating tasks to estimate expenses. Returns pricing_url and cost breakdown.'
• Add tool annotations via server.tool() options: readOnlyHint: true and idempotentHint: true for both tools. This signals to agents that these calls are safe to retry and do not modify state.
• Add error recovery guidance in descriptions. E.g., 'If task not found, verify task_id is correct and the task was created on the specified endpoint. Call list_tasks() to find active tasks.'