0

Agentharm RL Env (Prime Intellect)

Fresh

AgentHarm environment to evaluate agentic reasoning and safety

Type
RL Env
Publisher
Prime Intellect
Capabilities
Safety
Tags
Tool UseAgent
Runtime
multi-turn
License
unknown
Size
v0.1.0
Published
Nov 2025
Canonical
app.primeintellect.ai/dashboard/environments/primeintellect/agentharm

Cite

Notes

Only stored in your browser.

Attribution

Attribution policy →

Evals this tool implements

1

Same problem set, this tool's harness. Run it to score a model on the test.

AgentHarm: Harmfulness Potential in AI Agents

UK AI Security Institute (UK AISI)

Assesses whether AI agents might engage in harmful activities by testing their responses to malicious prompts in areas like cybercrime, harassment, and fraud, aiming to ensure safe behavior.

ActiveSafeguards