Westhawaiitoday Critical Flaw in Major AI Platforms Exposes Security Vulnerabilities
Article Content
- •CoreBreak flaw affects major AI platforms: AWS, Google, and Vercel.
- •Critical CVE-2026-18236 (CVSS 9.3) patched by Google.
- •Out-of-band policy enforcement recommended to prevent future exploits.
At Black Hat 2026, researchers Aviyam Ivgi and Hedi Ingber revealed a significant flaw, named CoreBreak, affecting AWS Bedrock AgentCore, Google’s Agent Development Kit, and Vercel’s AI SDK. The flaw allows attackers to bypass security guardrails by manipulating message histories and approval mechanisms. AWS patched CVE-2026-18830 (CVSS 8.1), Google addressed CVE-2026-18236 (CVSS 9.3), and Vercel patched CVEs 2026-64650 and 2026-64651 (both CVSS 6.3). The researchers emphasized the need for out-of-band policy enforcement to prevent such vulnerabilities. CoreBreak also exposed risks of credential theft through injected instructions. The flaws highlight the inadequacies of current AI security measures, with agents relying solely on prompt rules failing in over half of the trials conducted by Redpanda. The urgency for improved security measures is underscored by the critical nature of the vulnerabilities discovered.
Ask AI about this cluster
Answers cite the sources they use
Timeline
More articles in this cluster (5)
Following this threat?
Track CVE-2026-18236 in your own feed — you're alerted when they show up in new reporting, leak sites or exploitation.
Free account · no card needed
Common questions
Which AI platforms are affected?
What are the CVSS scores of the vulnerabilities?
What should organizations do to mitigate this risk?
Continue Reading
Critical Zero-Day Exploited in BeyondTrust Remote Support by China-linked Group A critical vulnerability in BeyondTrust Remote Support, tracked as CVE-2026-102764 with a CVSS score of 9.3, has been exploited by the China-linked espionage group Armory Castle since August 29, 2026. This flaw allows unauthenticated remote attackers to perform file operations, gaining root control and access to…
Silent Lynx Targets CIS Governments via WinRAR Exploit The threat actor Silent Lynx has launched a series of attacks against government organizations in the Commonwealth of Independent States (CIS) from June 2025 to August 2026. These attacks exploit a vulnerability in WinRAR (CVE-2025-8088) to deploy reverse shells, including FoalShell and PickleShell, facilitating…