Engineering
Train on Grok Bot, then graduate the job to a Cursor cloud agent via Notion
The biggest time-savers on Grok Bot are also the spendiest — you need a graduate path off the burn.
What hurt
Martin Devon (@MartinDevon) is writing agents in Cursor to offload Grok Bot usage. Usual workflow: train @bot to do something, then optimize it to work on a Cursor cloud agent, passing info via Notion when needed. Tip only — no invented prompt. Not affiliated with xAI.
What to install or create
Prove the job on Grok Bot. Capture the method. Hand context via Notion. Re-run on Cursor cloud. Keep approval on irreversible steps.
Bots
Train-then-graduate desk
Proves a job on Grok Bot, writes the handoff into Notion, stages the Cursor re-run. Stops before merge/publish.
Plugins
Notion
Required
Files
Notion page with method, inputs, and done criteria for the graduated job.
Routines
None on the first run. Watch one manual pass first.
The prompt
This is a tip, not a pasteable system prompt. The author did not publish one. Follow the first-run steps.
First run
- 1Read https://x.com/MartinDevon/status/2101068779598049561. Tip only — do not invent a system prompt.
- 2Pick one recurring job that burns Grok Bot. Run it once on Grok Bot until the method is clear.
- 3Write the method + inputs into a Notion page the Cursor agent can read.
- 4Re-run the same job on a Cursor cloud agent using that Notion handoff.
- 5Compare outputs. Keep send, post, merge, and spend behind your yes.
Approval fence
Training runs may use Grok Bot. Shipping code, posting, and spending wait for your yes.
What to tweak
Steal train-then-graduate. Notion is the handoff tray, not the brain.
Source
Martin Devon @MartinDevon · Original post
Not affiliated with xAI. We did not write this prompt. Third-party template terms still apply.
Pro pack
Want the full setup, not just the idea?
Pro packs give you the job text, prompts, schedules and pitfalls for popular use-cases.