The vulnerabilities — including a 10-year-old RCE flaw — add to the recent spate of AI-discovered cracks in foundational ...