Atlas · GenAI 2026
AI Red Teaming
Prompt injection / jailbreaking defense
conceptPeak: 2024Red TeamingAI consensus: 3/3
Prerequisites
No prerequisites.
Recommended reference
OWASP (2025) 'LLM Top 10' — owasp.org/www-project-top-10-for-large-language-model-applications; definitive threat reference
Notes from AI deep research
Anthropic Opus
OWASP LLM Top 10 = biblia. Perez (2022) foundational paper automated red teaming
OpenAI Deep Research
Budowa zabezpieczeń [OA#69]
Google Deep Think
Złośliwe próby przełamania [G#83]
Related skills
- → is subcategory of: AI Risk Management(3/3)
- ← is part of: AI Toxicity Analysis(3/3)
- ← is subcategory of: Adversarial AI Testing(3/3)
- → is subcategory of: AI Safety(2/3)