Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Adjust current task timeout policy. Use when the task is expected to take long, or when the system is too aggressive switching models. Prefer increasing timeout for long-running tasks with steady progress.
| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-03 | ✗→✓ | ▲ Improved | -40% | 0% |
| case-17 | ✗→✓ | ▲ Improved | -53% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 332% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -68% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 18% | 0% |
动态调整当前任务的超时策略，主要用于避免"卡死检测"误触发。
| 参数 | 类型 | 必填 | 说明 | |-----|------|-----|------| | progress_timeout_seconds | integer | 是 | 无进展超时阈值（秒），建议 600~3600 | | hard_timeout_seconds | integer | 否 | 硬超时上限（秒，0=禁用） | | reason | string | 是 | 简要说明调整原因 |
长时间浏览器任务:
json{ "progress_timeout_seconds": 1800, "reason": "需要完成多步浏览器操作，预计耗时较长" }
该设置只影响当前会话正在执行的任务，不影响全局配置。
create-todo: 创建任务计划enable-thinking: 启用深度思考Other measured skills in the registry, with their headline benchmark lift.