Dimensions overall score 8.0

• An autoresearch-style pipeline powered by Claude Code discovers novel white-box adversarial attack algorithms that significantly outperform all existing (30+) methods in jailbreaking and prompt injection evaluations.method 90%
• The discovered algorithms achieve up to 40% attack success rate on CBRN queries against GPT-OSS-Safeguard-20B, compared to ≤10% for existing algorithms.result 95%
• The discovered algorithms generalize: attacks optimized on surrogate models transfer directly to held-out models, achieving 100% ASR against Meta-SecAlign-70B versus 56% for the best baseline.result 90%
• White-box adversarial red-teaming is particularly well-suited for automation using LLM agents because existing methods provide strong starting points and the optimization objective yields dense, quantitative feedback.technical 85%
• The research demonstrates that incremental safety and security research can be automated using LLM agents.technical 80%
• Automation in discovering adversarial attacks could be misused if not properly governed, presenting potential ethical concerns around AI security.limitation 90%
• Claudini replaces traditional manually designed adversarial attacks with AI-driven automated discovery, offering faster and more effective security solutions.technical 75%
• The market for AI security is growing, with major investments in safeguarding AI systems by big tech companies and financial institutions.market 70%

Competitive landscape

Competitor map is still being generated for this paper. Enable generation or check back soon.

Startup potential card