Grind Until Done: Long-Horizon Cloud Agents

Done means you can decide when grind mode fits and scope a long-horizon task for it.

Use Grind modeA setting that lets a cloud agent keep working autonomously for hours on long jobs like migrations or building something from scratch. when a migration or from-scratch build is too long for a single short run. Keep the boundary narrow.

Start small. Name the job, attach the context that proves the point and decide what evidence would make the output trustworthy.

Read the loop before touching the controls. The first beat frames the work, the second uses Cursor, the third checks the result and the fourth leaves a handoff someone else can inspect.

Treat this as a short practice loop, not a product tour. The task should be small enough that you can inspect the result without trusting the summary.

1. 1Pick a genuinely long task: a framework migration, a dependency move, or building something from scratch.
2. 2Scope it tightly with a clear definition of done the agent can check itself against.
3. 3Turn on grind-until-done and let the cloud agent work for hours, returning artifacts as proof.
4. 4Review the PR and artifacts; reserve grind mode for the long tail, not everyday tasks.

The exercise is complete only when the proof matches the requested outcome. If the proof is weak, reduce the scope or fix the context instead of adding more instructions.

Verification decides the next move.

Use the first failure signal as the next prompt. Broad retries usually make the run noisier; a narrow retry gives Cursor a concrete repair target.