Hack in your sleep
summary
Builds an autonomous red team that reasons like a human via finite‑state planning, tying tactic→technique→procedure to the MITRE ATT&CK matrix and importing procedures from atomic red team and local scripts. Demonstrates an Operator/Caldera-style control, deploys one mac agent, runs 11 procedures in under a second, parses command output into variables, stages files, compresses them and exfiltrates over HTTP (file.io example) successfully. Shows scheduling, goal predicates and jitter to run repeatable, realistic autonomous assessments.
AI-written summary · about
chapters (7)
- 0:00Start
- 6:45Decision logic: thinking in a linear way
- 9:22Understanding finite state machines (math)
- 14:28Adversarial decision making == human decision making
- 17:39Cyber Kill Chain
- 20:19How to hack computers
- 23:49Solving the challenge
related talks
12mIntroduction To Red Team Tooling And Opsec
40mFantastic Red-Team Attacks and How to Find Them
27mKeynote: Towards Building Safe and Secure Agentic AI
20mAI Red Teaming and Automation: Exploring Societal Risks in...
52mBuilding the Always-On Purple Team
52mCALDERA: Automating Adversary Emulation
48mMITRE ATT&CK: The Play at Home Edition
41mKeynote:Bend the Machine: The Hacker Path to Autonomy in the Real World