# Grue Lab

- **Event:** [The Harness Engineering & Model Wrangling Hackathon](https://cerebralvalley.ai/e/mongodb-nyc-hackathon)
- **When:** Sat, Sep 26 at 9:00 AM – 10:00 PM (EDT)
- **Where:** New York, NY
- **Team:** [Himali Saini](https://cerebralvalley.ai/u/himali15), [Joshua Huang](https://cerebralvalley.ai/u/jhuang314), [Elliott Kalt](https://cerebralvalley.ai/u/elkalt), [sebastian losada](https://cerebralvalley.ai/u/n-a)
- **GitHub:** https://github.com/JoshBong/zorkinator
- **Demo video:** https://www.loom.com/share/77e2d77299e3472d8f85febbad78484b
- **Gallery:** https://cerebralvalley.ai/e/mongodb-nyc-hackathon/hackathon/gallery
- **Page:** https://cerebralvalley.ai/e/mongodb-nyc-hackathon/hackathon/gallery/56

Zorkinator plays Zork I with a harness that turns every death into a lesson instead of forgetting it. It builds its own map and memory as it plays, and a rule only earns the right to block a risky move after it's proven, by mechanically replaying every past game, that it actually would have prevented a death. Everything it learns is stored and versioned in MongoDB Atlas, so you can watch the same model go from clueless to competent, one game at a time.

## More from The Harness Engineering & Model Wrangling Hackathon

- [MF Goalkeeper](https://cerebralvalley.ai/e/mongodb-nyc-hackathon/hackathon/gallery/53)
- [Alter](https://cerebralvalley.ai/e/mongodb-nyc-hackathon/hackathon/gallery/54)
- [44872](https://cerebralvalley.ai/e/mongodb-nyc-hackathon/hackathon/gallery/55)
- [OneShot(Aegis)](https://cerebralvalley.ai/e/mongodb-nyc-hackathon/hackathon/gallery/57)
- [CyclopsDiary](https://cerebralvalley.ai/e/mongodb-nyc-hackathon/hackathon/gallery/58)
- [Incident Memory](https://cerebralvalley.ai/e/mongodb-nyc-hackathon/hackathon/gallery/59)

---

Markdown version of https://cerebralvalley.ai/e/mongodb-nyc-hackathon/hackathon/gallery/56. Site index for agents: https://cerebralvalley.ai/llms.txt · full text: https://cerebralvalley.ai/llms-full.txt