We are a white hat hacker collective and frontier independent red team. We don't do safety research... we do danger research.
"How can our AI system actually be misused, and what are we going to do about it?"
BT6 answers with rapid exploration, meticulous cartography, unparalleled creativity, and grounded mitigations; not safety theater or lobotomizations.
When the "unbreakable" models need breaking, when the red teams run out of ideas, when the auditors say "looks secure to us": that's when we begin.
We're not just elite practitioners of the latest tooling and TTP's. We're the pioneers who invented them. Assembled from elite jailbreakers, red teamers, pentesters, bug bounty hunters, and security researchers who didn't just learn the playbook; they wrote it. Operators who've breached every frontier model and Fortune 500 perimeter they've touched.
"With great responsibility comes great power."
Fortes fortuna iuvat
AI & LLM exploitation. Prompt injection, jailbreaks, refusal erosion, policy contradictions, emergent unsafe behaviors.
Data & infrastructure. RAG poisoning, ingestion manipulation, traditional cyber, network penetration, supply chain compromise.
Social engineering. Digital and physical. Compliance drift, persona manipulation, on-prem infiltration, human-layer exploits.
Embodied & physical AI. Robotics security, agentic systems, real-world actuators, long-horizon breakdown, toolchain escalation.
27 Active Operators
BT6 missions are limited and prioritized for systems where AI failure creates material impact. We work with frontier labs, enterprises, and governments.