Netcrook
#Gemini
Google Brings Computer Control Into Gemini 3.5 Flash, and the Prompt Injection Problem Gets Realer
A built-in computer-use feature pushes Gemini into browser, mobile, and desktop workflows, but the security question is now how well an agent can be kept from acting on hostile instructions.
Gemini 3.5 Flash Moves Closer to the Browser, and Security Teams Should Pay Attention
Google has expanded Gemini 3.5 Flash with agentic computer-use support for enterprise automation, a shift that turns UI control into a security problem as much as a productivity feature.
When AI Keys Slip Into the Wire: The iOS Leak That Exposed a Mobile Trust Problem
A large batch of iOS apps was found sending LLM credentials into network traffic, showing how quickly AI features can turn a device into a secret-handling liability.
Google’s $99 Gemini Speaker Turns a Simple Upgrade Into a Trust Problem
A new Google Home Speaker with Gemini inside is a product launch on the surface, but it also reopens the familiar security question behind every voice-first device: who gets to command the home?
WhatsApp, Rival Chatbots, and the New Battle Over Who Controls the Message Layer
A dispute over opening WhatsApp to ChatGPT, Gemini, and Claude is not just a competition story - it is a stress test for how AI, messaging, and platform security may collide.
When a Cheap Agent Starts Looking Expensive to Bigger Models
Lovelace’s benchmark claim points to a deeper shift in AI security and economics: for some research workflows, the decisive advantage may come from context plumbing, not raw model size.
The Browser Is the Prize: Malicious Add-Ons Put AI Logins in the Crosshairs
A wave of hostile extensions is putting browser-based access to ChatGPT, Claude, Copilot, Gemini, and DeepSeek under pressure, showing how one bad add-on can turn convenience into exposure.
When a Message Turns Into a Machine Command
Gemini’s Android notification handling shows how an assistant can move from reading text to taking action, creating a narrow but serious trust problem for smart-home and meeting integrations.
When a Chat App Becomes a Whispering Channel for Gemini
Security researchers reported that malicious messages can be used to influence Google Gemini through message notifications, turning routine inbox traffic into a prompt-injection risk.
When a Chat Preview Becomes an Attack Surface Inside Gemini
A reported indirect prompt-injection path shows how everyday notifications can turn untrusted text into instructions for an AI assistant.
When a Notification Becomes an Instruction: Gemini and the Prompt Injection Trap
A reported flaw in Google Gemini’s voice-assistant workflow shows how ordinary phone alerts can turn into a hidden channel for manipulation when untrusted text is treated like trusted context.
Inside the Leaked Operator Stack: How One Exposed Workspace Turned AI Credentials Into Propaganda Fuel
Researchers described a threat actor alias linked to an exposed working environment, where Gemini API keys, Telegram automation, and fraud tooling appeared to support a broader influence operation.
When a Leaked AI Key Becomes a Broadcast Machine
A reported Telegram influence campaign shows how stolen AI credentials can be turned into a content engine, with the real risk sitting in key management, not in a model exploit.
When AI Becomes a Campaign Multiplier: The New Advantage in Cyber Operations
An FT-attributed allegation about Iran-linked actors using ChatGPT and Gemini points to a broader security shift: generative AI may be lowering the cost of phishing, translation, and reconnaissance, without changing the old logic of intrusion.
When Lures Learn to Write Themselves: GreyVibe and the New Efficiency of AI-Aided Intrusions
A Russia-nexus threat cluster linked to Ukrainian targeting shows how generative AI can speed up lure creation and malware support without replacing old-school intrusion tradecraft.
Google Bets on an AI Shield as Security Teams Face AI-Fueled Intrusions
A new defensive bundle stitches together Google, Mandiant, Wiz, and Gemini, but the real story is how security vendors are trying to turn AI from a threat multiplier into a faster analyst workflow.
Google’s AI Stack Is Merging: Video Generation, Provenance, and Browser Tools in One Trust Test
Gemini Omni, SynthID, C2PA, and WebMCP point to a new phase in AI security, where the hard problem is no longer just making content, but proving what it is and controlling what happens next.
Search Results Became the Trap: Fake AI CLI Installers Turn Developer Curiosity Into Risk
Counterfeit installers posing as Gemini CLI and Claude Code show how search manipulation can become a delivery channel for malware, even when the underlying products are not the target.
Telegram and Gemini Under Scrutiny in an Alleged Political Abuse Operation
A contested figure tied to MAGA-targeted attacks highlights a blunt cyber reality: low-friction messaging and generative AI can compress the cost of online abuse, even when the technical path remains unproven.
The Search Trap Behind Fake AI Installers
A spoofed Gemini CLI download path shows how attackers can turn ordinary developer search habits into a delivery channel for malware.