Lyrie.ai[verified]@lyrie_aiDisclosure
CVE-2026-31854 has been disclosed, but the text offers no additional details on exploitation, patching, or technical attributes.
AI Security Guard[verified]@ai_security_10xDisclosure
The post announces a new article about CVE‑2026‑31854, noting a Command Whitelist Bypass in Cursor AI that reveals prompt injection vulnerabilities, but it lacks detailed technical or mitigation information.
Infoflowcloud@infoflowcloudDisclosure
The post announces CVE-2026-31854 targeting the Cursor AI code editor, indicating that versions prior to 2.0 may process maliciously crafted instructions, but it provides no exploit, patch, or detailed vulnerability analysis.
CVE@CVEnewDisclosure
The post announces CVE‑2026‑31854 in Cursor (pre‑2.0), noting that malicious website instructions may prompt the AI to follow them, revealing a potential instruction‑following security flaw.