Security
CISA Orders 3-Day Patch for Actively Exploited Zimbra RCE Flaw
CISA mandates urgent patching of a Zimbra RCE bug, while new research unveils a stealthy malware loader, encrypted AI jailbreaks, and an Android trojan.
This edition was produced with artificial intelligence. Text and voice are generated automatically.
CISA Gives Agencies 3 Days to Patch Actively Exploited Zimbra Command Injection Flaw
CISA added the flaw to its Known Exploited Vulnerabilities (KEV) catalog on Friday and gave Federal Civilian Executive Branch (FCEB) agencies until August 24 to secure their systems. The vulnerability is triggered by specially crafted SMTP requests when SNMP notifications are enabled, leading to the execution of arbitrary OS commands as the Zimbra user. Shadowserver tracks over 12,000 Zimbra servers exposed on the internet. CERT Polska has advised security teams to check logs for suspicious activity, including unexpected Zimbra service restarts and files created in specific webapp and tmp folders over the last 30 days.
CISA orders urgent patching of actively exploited Zimbra flaw →
New WordlistLoader Malware Uses Plain English Words to Hide Amatera Infostealer
WordlistLoader is distributed via ClearFake campaigns, which use ClickFix-style social engineering to trick users into executing malicious commands. The loader also unhooks modules used by security products, bypasses Event Tracing for Windows, and employs anti-emulation tricks. Amatera, a malware-as-a-service strain based on ACR Stealer, has become increasingly prevalent and can sweep up credentials, cryptocurrency wallet data, and browser history. Researchers note that law enforcement actions against other popular stealers like StealC and Lumma Stealer are pushing threat actors toward alternatives like Amatera, making it a growing concern.
Foul Language: WordlistLoader Disguises Malware as Ordinary Text →
Encrypted Prompts Bypass Grok and Gemini Guardrails, Exfiltrating Chat Histories
The technique differs from earlier cipher-based jailbreaks by using strong authenticated encryption, which forces execution and launders the payload’s provenance. Against Gemini 3 Flash, the same approach produced instructions for building incendiary devices and caused the model to reproduce its own system instructions. Adversa reported the Grok issue to xAI on June 3 but received no substantive response despite follow-ups, and confirmed the attack still reproduced on August 19. The Gemini finding was never formally reported because Google’s vulnerability reward program excludes prompt injection and jailbreaks from scope, leaving such research without a coordinated-disclosure track. The findings highlight that content-based guardrails are structurally unable to inspect payloads that only exist in plaintext after the model has already decided to trust them.
Encrypted Prompts Defeat Grok and Gemini Guardrails; Chat Histories Stolen →
ToxicPanda 2.0 Android Trojan Uses Fake VPN Prompt to Disable Google Play Protect
With Google’s defenses disabled, ToxicPanda decrypts a hidden payload, installs it, and requests Accessibility Service permissions. The malware supports 167 remote commands and can overlay fake login screens on 349 banking, e-wallet, and crypto apps across 16 countries, a significant increase from 16 apps in its previous iteration. It also harvests PINs from over 140 financial apps using invisible overlays, can spoof the Android lock screen to steal credentials, and abuses Android’s Wireless Debugging (ADB) feature to gain shell-level access. ToxicPanda first appeared in 2024, targeting European banks.
This Android banking trojan uses a fake VPN prompt to silence Google’s defenses →
SickKids Hospital in Canada Hit by Second Cyberattack, Employee Data Exposed
An investigation found that personal information of some former and current employees of SickKids, Boomerang, and SickKids Foundation, as well as job applicants, was exposed. The nature of the exposed information and the number of affected individuals were not disclosed. Affected individuals have been offered 24 months of complimentary credit monitoring and identity protection services. SickKids stated it remains committed to maintaining strong protections and enhancing its cybersecurity measures.
Canadian SickKids hospital hit again by cyberattacks, more data stolen →