AI jailbreak severity framework
AssessTechniques
A proposed framework for categorizing how severely a jailbreak can bypass an AI model's safeguards.
Why it's here
Placed in Assess: 1 article(s) of evidence from 1 source(s), led by framework updates, with 0 in the last 30 days. Confidence 24%. Low accumulated evidence, so it defaults conservatively pending more signal.
Evidence (1)
- 7Anthropic News·7/2/2026framework_updateAnthropic details Fable 5 cyber safeguards and jailbreak severity framework
Anthropic says Claude Fable 5 has been redeployed globally and is now available to all users, alongside a fuller explanation of its cybersecurity safety classifiers. The company also introduced an early draft framework for grading AI jailbreak severity and opened a HackerOne program for researchers to report cyber jailbreaks.