すべてのユースケース

エンジニアリング

First mile, last mile: let Grok Bot find the work, then close the loop with Cursor

Most of the effort is deciding what to work on and then actually finishing it — the middle is the easy part.

共有

困っていたこと

Lauren (@poteto), who builds Grok Bot at SpaceXAI, says her favorite use is the first and last mile of work. First mile: connect your apps so the bot can pull its own context. Last mile: hand code fixes to Cursor cloud agents and close the loop. She shared the five-step loop she runs daily. Not affiliated with xAI.

インストールまたは作成するもの

Connect the places work shows up (Slack, CRM, email, calendar). Give context instead of micromanaging, like her example below. Add a routine on a schedule or on a new channel message. For code, the bot hands the job to a Cursor cloud agent. Once that feels solid, wire her loop: feedback → Linear ticket → reproduce → fix + fuzz → ping, then merge.

ボット

  • Main bot

    Watches the feedback channel, files the Linear ticket, and starts the Cursor cloud agent.

プラグイン

  • Linear

    任意

  • GitHub

    任意

ルーティン

  • Feedback watcher. When the feedback Slack channel gets a new message (author's setup). Ticket it, reproduce it, propose a fix, then ping you.

プロンプト

First-mile context example (author-quoted)

Quoted from @poteto — as published. Do not invent a second system prompt.

look at this slack channel for new features we've built and then look through our CRM to see which customers we should nudge

Daily feedback-to-fix loop (author-quoted)

Quoted from @poteto — as published. Do not invent a second system prompt.

1. watch this slack channel for feedback about our app
2. create a ticket for this in Linear
3. create a Cursor cloud agent to triage and reproduce the issue on its own computer, using pstack
4. if the issue clearly reproduces, put up a fix and fuzz the pull request with a small swarm of agents
5. if the fuzz has no issues, ping me on Slack, and rebase and auto-merge the PR after 1 hour unless i request changes

初回実行

  1. 1Read https://x.com/poteto/status/2107510472601985336 (credit @poteto).
  2. 2Connect the apps where your work shows up, and give the bot context instead of step-by-step orders (her example below).
  3. 3Add one routine: on a schedule, or whenever a chosen channel gets a new message.
  4. 4For code, have the bot hand the fix to a Cursor cloud agent instead of coding it itself.
  5. 5When that works, copy her five-step loop below — and keep the final merge manual until you trust it.

承認フェンス

Her loop auto-merges after an hour unless she objects. Start with the merge in your own hands, and only add a timed auto-merge once the reproduce-and-fuzz steps have earned your trust.

調整ポイント

Give the bot context and a trigger, not a script. Let it pick the work, and let a coding agent finish it.

出典

lauren @poteto · 元の投稿

xAI とは提携していません。このプロンプトは当サイトが書いたものではありません。第三者テンプレートの規約も適用されます。

テンプレートを見る

チーム設計図を見る

Pro パック

アイデアだけでなく、完全な構成が必要ですか?

Pro パックには、人気ユースケースのジョブ本文、プロンプト、スケジュール、落とし穴が入っています。

Pro を見る