Six curl CVEs after OpenAI and Anthropic came back with zero

AISLE's autonomous AI system discovered six CVEs in curl within days of OpenAI Codex Security and Anthropic Mythos reporting zero vulnerabilities in the widely-deployed software. All six vulnerabilities were rated low severity and were fixed in curl 8.22.0, demonstrating that frontier AI systems from major companies failed to detect flaws that AISLE found, suggesting potential limitations in existing AI-based cybersecurity approaches.