Execution accountability for AI agents.
We're building in public. Open-source releases land here when they're ready — not before.
| Product | Status | What it does |
|---|---|---|
| AgentBench OS | 🔜 Coming soon | CI gates for agent regressions on pull requests |
| Witness | 🔜 Planned | Replay and bisect agent execution runs |
| K8sAttest | 🔜 Planned | Proof-carrying Kubernetes deploys from agent-generated configs |
No product repositories are public yet. Watch this org for the first ship.
Casualstack builds developer infrastructure so AI coding agents can't silently break your codebase or your cluster — measure, replay, and prove correctness before merge.
Stack: Python · GitHub Actions · property-based oracles · attestation
- Website: casualstack.dev
- Discussions: opening when the first repo ships
Building in public.