AI Red Team Simulation

İlerisecurityEn az 32K bağlam

Simulates realistic adversarial behavior against AI features to uncover prompt injection, data exfiltration, unsafe tool use, jailbreak, and authorization failures before launch.

red-teamai-safetyprompt-injectionthreat-modeling

Kullanım alanları

  • AI security reviews
  • Agent threat modeling
  • Pre-release adversarial testing

Örnek prompt

Create a red-team plan for this AI workflow. Model attacker goals, trust boundaries, adversarial prompts, tool abuse paths, expected controls, and reproducible test cases.

Önerilen modeller

claude-opus-4-6gpt-5gemini-2-5-pro

Uyumlu araçlar

claude-codecursorkiroany

Modaliteler

Giriş: text, code, file
→
Çıkış: text, code

Yazar

OpenModels Community

@openmodelsrun