AI Vulnerability
-
Drupal warns of critical security update later today
Drupal plans a core security update later today and warned that exploits could follow within hours. The advisory affects core versions 8 and later, with fixes planned for supported releases and hotfixes for some older branches.
-
Researchers find flaw that could let websites inject prompts into Anthropic’s Claude Chrome extension
Researchers disclosed a flaw called ShadowPrompt in Anthropic’s Claude Chrome extension that combined an overly permissive origin allowlist and a DOM-based XSS in an Arkose Labs CAPTCHA, allowing websites to inject prompts; Anthropic and Arkose issued fixes in December 2025 and February 2026.
-
Critical vulnerability in n8n workflow platform could allow code execution (CVE-2025-68613)
A critical vulnerability in the n8n workflow automation platform (CVE-2025-68613) with a CVSS score of 9.9 could allow authenticated users to trigger arbitrary code execution; patches are available and Censys reports over 103,000 potentially vulnerable instances.
-
CISA adds critical ASUS Live Update flaw to known exploited vulnerabilities catalog
CISA added a critical ASUS Live Update vulnerability, CVE-2025-59374 (CVSS 9.3), to its Known Exploited Vulnerabilities catalog citing active exploitation; the flaw stems from a past supply chain compromise and vendors say affected builds were limited to devices meeting specific targeting conditions.
-
Fortinet warns of FortiWeb OS command injection flaw CVE-2025-58034 exploited in the wild
Fortinet warned that CVE-2025-58034, a medium-severity OS command injection in FortiWeb with a CVSS score of 6.7, has been exploited in the wild; patches are available in specific FortiWeb releases and the company credited a Trend Micro researcher for the report.
-
Amazon opens invite-only bug bounty for NOVA models to outside researchers
Amazon has launched an invite-only bug bounty program for its NOVA family of language models, allowing select researchers to test and be paid for findings on issues such as prompt injection, jailbreaking and other vulnerabilities, with the company saying the effort will help secure models integrated across Amazon and customer systems.
-
TP-Link issues firmware updates for Omada gateways to fix four vulnerabilities, including two critical bugs
TP-Link released firmware updates for Omada gateway devices to fix four vulnerabilities, including two critical command injection flaws; users are advised to apply updates and verify device configurations.
-
Google DeepMind unveils CodeMender to detect, patch and rewrite vulnerable code
DeepMind has unveiled CodeMender, an AI agent that detects, patches and rewrites vulnerable code using Gemini models and an LLM-based critique tool; Google says it has upstreamed 72 fixes and is expanding AI security measures including an AI Vulnerability Reward Program and updates to its Secure AI Framework.
-
Security firm flags in-the-wild exploitation of Pandoc flaw CVE-2025-51591 to target AWS IMDS
Security researchers at Wiz have observed in-the-wild exploitation of CVE-2025-51591, a Pandoc flaw that enables SSRF against AWS EC2 IMDS, with attackers attempting to exfiltrate data via crafted iframes. The activity underscores the importance of IMDSv2 and least-privilege IAM roles to mitigate cloud credential exposure.
-
Cursor AI editor vulnerability could enable covert code execution on folder open, researchers warn
A vulnerability in Cursor, the AI-augmented fork of Visual Studio Code, could allow attackers to silently run code on a user’s machine when a repository is opened, researchers warn, due to default Workspace Trust settings and potential autorun configurations.
