# AGET

- **Event:** [Claude Build Day](https://cerebralvalley.ai/e/claude-startups-build-day)
- **When:** Sat, Jun 13 at 9:00 AM – 10:00 PM (PDT)
- **Where:** San Francisco, CA
- **Team:** [Gabor Melli](https://cerebralvalley.ai/u/gmelli)
- **GitHub:** https://github.com/gmelli/aget-bench
- **Demo video:** https://youtu.be/Qs9mzGoKyA0
- **Gallery:** https://cerebralvalley.ai/e/claude-startups-build-day/hackathon/gallery
- **Page:** https://cerebralvalley.ai/e/claude-startups-build-day/hackathon/gallery/83

AGET-Bench — a benchmark for governance-overlay agent behavior. This public slice asks: does a coding model's rule-compliance degrade as more rules are presented to it? Compliance is computed by a deterministic rule-checker executed against the model's generated code — never self-reported. Result: density-invariant for Opus 4.8 — compliance holds across a 4× range of rule count (5 / 10 / 20 rules: 80% / 75% / 76%, 95% CIs overlap). Build Day is the public release of this ongoing benchmark, with transparent handling of degenerate (prose-only) runs.

---

Markdown version of https://cerebralvalley.ai/e/claude-startups-build-day/hackathon/gallery/83. Site index for agents: https://cerebralvalley.ai/llms.txt · full text: https://cerebralvalley.ai/llms-full.txt