Weft
繁
任務 · MCP

Ask the User a Question

request_input is one of the 32 MCP tools in Weft, the AI task manager you run by talking to Claude, Cursor, Codex or ChatGPT through one MCP endpoint. Ask the person something only they can decide, and wait without losing the claim on the task. If they are at their client, a form appears there and the answer comes back inside the same call — the agent carries on in the same context, with the same claim. If they are away, the call returns parked and the question waits in their queue; the answer reaches the agent at its next start_task or get_my_work. For the irreversible and the genuinely ambiguous, not for "may I continue?" — every question spends the same attention budget that limits new work.

Weft 是為獨立創業者和小團隊打造的 AI 任務管理工具。

工具參考文件以英文維護：工具名稱、參數與範例在每種語言中都相同。

參數

名稱型別必填說明
idstring是Task ID
runIdstring是The runId returned by start_task
leaseTokenstring是The leaseToken returned by start_task
questionstring是One sentence a person can answer (≤500 chars)
options{ value, label }[]否The alternatives, when there are any. A question with options gets read; a yes/no gets rubber-stamped.
urgency'blocking' | 'soon'否Default 'blocking'
holdSecondsnumber否How long to wait inside the call before parking (default 120, max 240)

回傳

Either answered:true with the answer and the same runId/leaseToken to continue with, or parked:true with a questionId — the claim is kept either way, and a parked question is delivered back through start_task and get_my_work.

註記

  • 冪等。用相同參數呼叫兩次，效果和呼叫一次一樣。

範例

使用者說的話：

"Ask me before you send anything to a customer."

AI 在背後呼叫的東西：

json
{
  "name": "request_input",
  "arguments": {
    "id": "…",
    "runId": "…",
    "leaseToken": "…",
    "question": "Which discount should the email offer?",
    "options": [
      {
        "value": "10",
        "label": "10%"
      },
      {
        "value": "15",
        "label": "15%"
      }
    ]
  }
}

相關