Prime Environments (safety envs)

Evals

Environment / sim

Community RL envs; includes agentharm, agentic_misalignment, harmeval, jailbreak_bench, wildjailbreak.

Category
Eval frameworks & harnesses
Insurance exposure
Agent actions
Publisher
Prime Intellect
Publisher type
Neo-lab
License
Apache-2.0
Status
Maintained (<12mo)
Last commit
2026-09-07
Stars
263

View source on GitHub →

Wanich Risk Lab, 2026.