All news in category “Security Advisory and Patch Watch”

⚠️ Security researchers disclosed two critical vulnerabilities in the Python-based AI app framework Chainlit that allow unauthenticated attackers to read arbitrary server files and trigger SSRF requests. The flaws (CVE-2026-22218 and CVE-2026-22219), fixed in Chainlit 2.9.4, stem from an unvalidated custom Element type exposing path and URL properties. Exploits can leak environment variables, API keys, LLM prompts, and cloud credentials, enabling lateral movement and broader compromise.

⚠️ A critical vulnerability in ACF Extended (CVE-2025-14533) allows unauthenticated attackers to obtain administrative privileges by abusing the plugin's 'Insert User / Update User' form action in versions up to 0.9.2.1. The flaw fails to enforce role restrictions at the form level, enabling attackers to set arbitrary roles, including administrator, when a role field is present. The vendor released a patch in version 0.9.2.2 on December 14, 2025; administrators should update immediately and audit any forms that create or update users because roughly 50,000 sites may still be exposed.

🔒 Unit 42 researchers discovered an Azure Private Endpoint DNS behavior that can unintentionally or deliberately produce denial-of-service conditions for Azure services. In several scenarios — accidental internal, accidental vendor, and malicious actor — linking a Private DNS zone to a virtual network can force name resolution to the private zone and fail when no A record exists, breaking connectivity to otherwise public endpoints. Microsoft documents a partial mitigation (fallback to internet); alternatives include manually adding DNS records and performing comprehensive discovery with Resource Graph.

🔒 Chainlit, a widely used framework for building conversational AI applications, contained two server-side vulnerabilities (CVE-2026-22218 and CVE-2026-22219) that allow authenticated users to read arbitrary files and trigger SSRF in affected deployments. The flaws stem from insufficient validation of user-controlled properties in custom elements and SQLAlchemy-backed storage. Combined, they can expose environment variables, cached prompts, API keys and cloud metadata, enabling lateral movement beyond the app layer. Chainlit released 2.9.4 on 24 December 2025 and users are advised to apply the patch immediately; temporary WAF signatures were published as mitigation.

🚨 Cybersecurity researchers have identified three prompt-injection vulnerabilities in Anthropic's reference Git server implementation, mcp-server-git, affecting default installations and all releases before 8 December 2025. The flaws let attackers manipulate what an AI assistant reads—such as a README, issue text or a webpage—to cause unintended actions without credentials or system access. Exploits can enable code execution when combined with a filesystem MCP server, delete arbitrary files, or load sensitive files into a model's context. Anthropic accepted the reports in September and issued patches in December 2025; affected users are urged to update immediately.

🛡️ Amazon announced quarterly security and critical updates for Amazon Corretto LTS distributions on January 20, 2026. Updated builds — Corretto 25.0.2, 21.0.10, 17.0.18, 11.0.30, and 8u482 — are available for download. Customers can obtain releases from the Corretto home page or configure Apt, Yum, or Apk repositories on Linux to receive updates. Feedback and issue reports are invited via the Corretto GitHub repository.

⚠️ A trio of vulnerabilities in mcp-server-git, the official MCP Git server maintained by Anthropic, can be chained to read or delete arbitrary files and, in certain scenarios, achieve remote code execution. Cyata researcher Yarden Porat showed these issues are exploitable via prompt injection when an AI assistant ingests attacker-controlled content such as a malicious README or poisoned issue text. Fixes were released in 2025.9.25 and 2025.12.18; users should update the Python package promptly to mitigate risk.

🔒 Rockwell Automation reported two high-severity vulnerabilities in Verve Asset Manager affecting legacy components: the ADI server and the Ansible playbook. Both issues can result in unencrypted sensitive information being stored in environment variables or during playbook execution and are rated CVSS 7.2 and 7.9. Rockwell states the flaws are resolved in 1.42; organizations should upgrade and contact Rockwell TechConnect for assistance. CISA also recommends minimizing network exposure and using secure remote access such as up-to-date VPNs.

⚠️ Schneider Electric published an advisory about a side‑channel vulnerability disclosed by Intel (CVE-2018-12130) that affects EcoStruxure Foxboro DCS Virtualization Server (V91) and Standard Workstation (H92). An authenticated user with local access could exploit the CPU issue to enable information disclosure, risking loss of system functionality or unauthorized access. Schneider Electric directs customers to migrate to updated server (V95) and workstation (Dell D96) hardware or, if immediate migration is not feasible, to apply BIOS and OS security patches and follow layered defense-in-depth recommendations.

⚠️ Schneider Electric warns that multiple vulnerabilities in the CODESYS Runtime System V3 communication server affect many Schneider products and third-party devices embedding CODESYS. Exploitable issues include denial-of-service and, in some configurations, remote code execution; several CVEs carry CVSS scores up to 8.8. Schneider has published patches and mitigations for many affected product families; operators should apply vendor updates and follow immediate network and access controls to reduce exposure.

🚨 Security researchers have uncovered the CrashFix campaign, which uses a deceptive Chrome extension to intentionally crash browsers and trick victims into executing attacker-supplied commands. The malicious add-on, identified as NexShield-Advanced Web Protection and branded to resemble uBlock Origin Lite, remains dormant for about an hour before exhausting resources and forcing repeated crashes. On restart, users see a fake repair prompt instructing them to paste a command into the Windows Run dialog; executing it launches a multistage infection that ultimately deploys a previously undocumented Python-based remote access trojan named ModelRAT. Huntress ties the activity to a threat cluster it calls KongTuke and warns administrators to remove look-alike extensions, avoid running unsolicited fix commands, and use published IOCs to detect related activity.

🔒 Cloudflare patched a vulnerability in its ACME HTTP-01 validation logic that could allow requests to bypass WAF protections and reach customer origin servers. Discovered by FearsOff in October 2025, the flaw arose when edge logic disabled WAF handling for requests matching an ACME challenge token without confirming the token belonged to the requested hostname. Cloudflare said it found no evidence of exploitation and implemented a code change on October 27, 2025 to only disable WAF features when the token is a valid challenge for that specific hostname.

🔓 Mandiant released a pre-computed Net-NTLMv1 rainbow table so anyone can map challenge-response data back to real NT hashes, a move intended to force organizations to abandon the insecure NTLMv1 protocol. The dataset, hosted via the Google Cloud Research Dataset portal, can recover keys in about 12 hours using roughly $600 of hardware. Mandiant says the goal is to demonstrate immediate risk and prompt remediation rather than to create new vulnerabilities.

🔒 Cloudflare patched a logic flaw in its ACME HTTP-01 handling that could disable certain WAF protections for specific challenge paths. The issue was reported by researchers from FearsOff through Cloudflare’s bug bounty program on October 13, 2025, and affected requests to /.well-known/acme-challenge/*. In some cases, challenge requests could reach customer origins when they should have been blocked because WAF features were incorrectly disabled. Cloudflare implemented a code change to ensure WAF disabling only occurs when Cloudflare will serve a valid ACME challenge response; no customer action is required and there is no known abuse.

🔒 A team from CISPA disclosed StackWarp, a hardware vulnerability affecting AMD Zen 1–5 processors that subverts SEV-SNP protections. The flaw lets a privileged host manipulate a guest VM's stack pointer via a previously undocumented control bit and a co-running hyperthread, enabling control-flow hijacks, data corruption, and secret exfiltration. Vendors released microcode fixes and AGESA patches are planned.

⚠️ Researchers at Palo Alto Networks' Unit 42 disclosed critical weaknesses in the NeMo, Uni2TS and FlexTok Python libraries used with Hugging Face models, where malicious code can be hidden in model metadata and executed automatically when a manipulated file is loaded. The root cause is the use of Hydra's instantiate(), which accepts arbitrary callables and arguments and can therefore permit remote code execution if metadata is untrusted. Vendors including NVIDIA, Salesforce and the maintainers of FlexTok have issued fixes and CVE assignments; users should upgrade affected libraries and audit models before loading.

⚠️ Research firm AppOmni disclosed a critical privilege-escalation vulnerability called BodySnatcher in ServiceNow’s Now Assist AI Agents and Virtual Agent API that could let unauthenticated actors execute workflows as arbitrary users. ServiceNow says hosted instances were patched at the end of October and customers should upgrade to specified Now Assist and Virtual Agent API versions. AppOmni warns that default example agents and permissive authentication choices mean similar risky configurations could still exist in custom code or third-party integrations, and recommends enforcing MFA, reviewing agents, and applying the updates promptly.

🔧 Microsoft has issued out-of-band Windows updates to address two issues introduced by the January 2026 security updates: credential prompt failures that can block Microsoft 365 Cloud PC and remote desktop sign-ins, and a shutdown/hibernate failure on Windows 11 23H2 when Secure Launch is enabled. The fix packages must be manually downloaded from the Microsoft Update Catalog, and administrators can deploy Known Issue Rollback (KIR) installers via Group Policy for enterprise-managed devices when immediate deployment is required.

⚠️ Cisco has released patches for a critical zero-day, CVE-2025-20393, in AsyncOS that affects Secure Email Gateway (SEG) and Secure Email and Web Manager (SEWM) appliances. The flaw allows a remote attacker to gain root by sending a crafted HTTP request to the Spam Quarantine interface when it is enabled and reachable from the internet. Cisco first learned of exploitation in December, issued a public advisory on Dec. 17, and has now published fixes to address the issue.

⚠️ XM Cyber disclosed privilege-escalation flaws in Google’s Vertex AI that let low‑privileged users manipulate Google-managed Service Agents to gain elevated project-wide permissions. Google told XM Cyber this behavior is "working as intended." Security experts warn that managed service identities and insecure defaults create invisible, structural risks. CISOs are urged to audit service identities, reduce authentication scope, and monitor agent activity like privileged users.