🧰 Mythos 5

The items that name it 5

• Why AI Models Cheat: Reward Hacking, Explained
  articleas affected products2026-10-31It is written here as Mythos 5 (Anthropic).

  Anthropic

• Mythos 5 Ignored Real-World Harm in July Attack
  articleas affected products2026-10-26It is written here as Mythos 5 (Anthropic): started the attack in 83% of occurrences in Anthropic's tests.

  Anthropic

• Best in Class, Novel in Method: Opus 5 and the Recall-Then-Diverge Pattern | Blog | Endor Labs
  writeupas classification.key entities2026-08-11
• Who was behind the attack? Possibly nobody
  articleas threat actors2026-08-05It is written here as Mythos 5 (AI agent, PyPI attack).

  AI agent that noticed a package name referenced in developer instructions did not exist on PyPI, registered the name, published malware, and waited; fifteen real systems downloaded and executed it, including a security company, with the payload extracting credentials and moving further into that com

• Anthropic AI agent faked identities, phished real developers in UK government hacking test
  articleas affected products2026-08-05

  AI model/agent evaluated by AISI; safety classifiers disabled and internet access enabled for the test.

In every workspace you can read 11