AI Signal 157
AISLE found six curl CVEs after OpenAI and Anthropic reported zero findings
↗
AISLE's autonomous AI system discovered six confirmed CVEs in curl 8.22.0, days after OpenAI Codex Security and Anthropic Mythos both reported zero findings in the same codebase.
This demonstrates that specialized AI security systems can outperform frontier AI lab products at real-world zero-day discovery, even on heavily audited codebases like curl deployed across more than 20 billion instances. The Linux stable maintainer reports seeing the same pattern, suggesting this is not isolated to one project.
Written by elseif from the cluster below · every claim links back to a sourceThe three things worth knowing
AISLE found six confirmed CVEs in curl after OpenAI and Anthropic's frontier AI systems both reported zero findings.
All six vulnerabilities are rated Low severity, consistent with curl's maturity where remaining flaws hide in narrow configurations and subtle interactions.
Linux stable maintainer Greg Kroah-Hartman reports AISLE is finding similar results in the Linux kernel.
THE CLUSTER