MITRE ATLAS
MITRE ATLAS
Published by The MITRE Corporation · Current live matrix on atlas.mitre.org
A globally accessible, living knowledge base of adversary tactics and techniques involving AI — including attacks against AI-enabled systems, abuse or manipulation of AI capabilities, and harmful autonomous behavior materially enabled by AI — based on real-world observations and realistic red-team demonstrations. Modeled after and complementary to MITRE ATT&CK, with several tactics and techniques adapted directly from the ATT&CK Enterprise matrix.
How it’s structured
- Organized as a matrix of tactics (the adversary's goal, the "why") and techniques (how that goal is achieved, the "how") targeting AI-enabled systems
- Currently lists 16 tactics, 208 techniques, 40 mitigations, and 73 case studies
Who it applies to
Defenders, red teams, and owners of AI-enabled systems who want to map threats and defenses against real-world attacks on AI/ML systems.
How Pentforge uses it in an engagement
- Threat mapping for AI-agent and LLM-application engagements, scoped to the ATLAS tactics and techniques relevant to the client's deployment
- Findings referenced back to the specific ATLAS technique for a shared vocabulary with the client's team
What you get
- A findings report mapped to specific ATLAS techniques
- A coverage summary showing which AI-specific attack techniques were exercised
Official source
Pentforge does not issue certificates
We assess against MITRE ATLAS — that is a statement about the work, not a certification. Pentforge holds no certifications of its own; the credentials our practitioners hold individually are listed on About.
Request a scoping call