Evil Agent Dojo Be Like

Fresh

Evil AgentDojo: Training models to be MORE susceptible to prompt injections (inverted reward signal)

Type: RL Env
Tags: Security Prompt Injection Adversarial AI Safety Misalignment
License: unknown
Size: v0.1.3
Published: Nov 2025
Updated: Nov 2025
Canonical: app.primeintellect.ai/dashboard/environments/skdrx/evil-agent-dojo-be-like

Cite

Notes

Only stored in your browser.

Attribution

README: api.primeintellect.ai/api/v1/environmentshub/skdrx/evil-agent-dojo-be-like/@0.1.3/inspect

Attribution policy →

Lift evidence

1

Eval	Tools known to lift	Source paper
AgentDojo: A Dynamic Environment to Evaluate Prompt Injection Attacks and Defenses for LLM Agents	Evil Agent Dojo Be Like	-

Contributors

1