How the Hack Works
You have two hours to turn a real work problem into a working agent.
You pick your altitude, pick a scenario, and demo what you made. Coaches are in the room throughout. Aim for something that works and that you understand well enough to keep using, not something polished.
The flow
- Pick your altitude - Cowork with no code, Scout in between, or real code in VS Code.
- Pick a scenario - a real Microsoft work problem to point it at.
- Choose your data path - Cowork and Scout read your own work privately; Code ships a fictional twin so no personal data is involved.
- Build the smallest working version - get to v0.1 fast.
- Test it on a real prompt - see where it breaks.
- Patch the spec - change the rule, sample, or boundary that caused the miss.
- Demo the behavior - show the before, the patch, and the after.
Ground rules
- Nothing sends. Your agent drafts for human review. It does not send email, book meetings, or make commitments on your behalf.
- Keep private data private. If you use your own Microsoft 365 data, keep it on your own screen.
- Build small. A short spec outperforms a long one.
- Use observed behavior. What you actually did last time matters more than what you wish you did.
- Portability matters. Prefer files and formats you can carry across tools.
What "done" looks like
Done means you can run your agent on a fresh prompt and get useful output that reflects the scenario.
What the artifact is depends on where you landed - a portable persona spec, a room of teams that routes a request into a plan, an app grounded in real data. The shape differs; the bar is the same: it runs, it's specific, and you can point to the rule behind its behavior.
What you'll walk out with
You leave with two things:
- A working artifact you can keep improving after the hack.
- A transferable skill: how to compress messy work behavior into a small, testable agent spec.
How you're judged
It's not a competition with points - the aim is a working agent you understand. See how you're judged →