ControlArena

Evals

Framework / harness (not an eval — tooling/reference entry)

Library of control settings (BashArena, Apps, infra) to test monitors vs scheming agents.

Category
AI control & monitoring
Insurance exposure
Agent actions
Publisher
UK AISI + Redwood Research
Publisher type
Government / standards body
License
MIT
Status
Active (<90d)
Last commit
2026-10-13
Stars
253

View source on GitHub →


Wanich Risk Lab, 2026.