🎯 Red teaming

The items that name it 6

• GPT-Rosalind-6 System Card Addendum
  paper2026-10-30

  A biosecurity red-teaming effort was run on GPT-Rosalind-5.5 to test monitoring and enforcement, and is being repeated for GPT-Rosalind-6; details were shared with the Safety Advisory Group but not published.

• Part 3 – Cybersecurity journeys: The work behind the work
  article2026-10-27It is written here as Red teaming: offensive security work Kurt Heath is working toward.
• zero-lab-ai/skills
  repo2026-10-11
• Anthropic Introduces 3-Tier Cyber Verification Program for AI Access
  article2026-10-07

  Authorized red teaming is allowed in the Red Team Access tier.

• AI Security Briefing
  article2026-07-18

  Cybersecurity practice of deliberately attempting to break a system to identify weaknesses before attackers can exploit them; applied by both human researchers and AI systems like GPT-Red

• Orca Security: NIST AI Risk Management Framework (AI RMF) Explained
  article2026-05-20

  Adversarial testing practice referenced as part of trustworthy AI control practices