Zero-Click Grok Chat History Theft: Adversa AI Demonstrates Cryptographic Context Injection
2026-08-23T08:51:36Z•e351eb27223a10023f07ca95640140043f7a0bfa72cba20716619bd46fdee645
CVE-2026-19478CVE-2026-73570AI securityAndroid malwareBADBOXCISA KEVCiscoEMVGitLabGrokNASA JPLOAuth abuseRussian espionageToxicPandaWhatsAppZimbraactive exploitationautomotive infotainmentdata exfiltrationfinancial institutionsmobile banking trojanpayment securityphishingprompt injectionspace systemsunauthenticated remote code execution
What happened
Security news digest covering zero-click cryptographic prompt injection against Grok chat histories, Android banking malware and BADBOX proxy infections, a critical unauthenticated command-execution flaw in NASA/JPL AIT-GUI, actively exploited Zimbra and GitLab vulnerabilities, Cisco maximum-severity flaws, Russia-linked phishing and OAuth espionage campaigns, and payment-card protocol weaknesses.
Why it matters
A reviewed impact interpretation has not been published for this record.
Evidence and limitations
- Source ID
- securityaffairs
- Record identifier
- e351eb27223a10023f07ca95640140043f7a0bfa72cba20716619bd46fdee645
- Enrichment time
- 2026-08-23T08:51:36Z
- AI-assisted enrichment
- Yes
This record may overlap with other records. Its enrichment can be incomplete or wrong, and machine assistance was used. Validate consequential decisions against the linked source and your own environment.