AI Red Teaming — Tools

LLMmap

Fingerprinting tool used to identify which model is actually powering a target AI application, without relying on the application (or the model) to self-report it.

• Use during recon/fingerprinting, as a follow-up to the Testing Methodology recon checklist — run it after the manual "model identity" questions to confirm or contradict what the model claimed about itself.
• Why it matters: self-reported model identity can be wrong or deliberately obscured (system prompt instructing the model to lie about what it is); a fingerprinting approach based on observed behavior is harder to spoof.

Stub

The source notes only say "run LLMmap" — no command usage, options, or output details were captured yet. Fill in with actual usage/output once run against a real target.