Researchers demonstrated a technique that can trick xAI’s Grok into leaking a user’s chat prompts and some session details to an attacker-controlled server when the user asks Grok to summarize a web page. The attack hides malicious instructions inside encrypted content on the page, which Grok is induced to decrypt and then use to build a data-leaking URL request, without an obvious warning or confirmation step in the proof-of-concept.
Mirage safely runs attacks like this one against your own team, so you find out what happens before a real adversary does.
You paste a link into Grok and say, “Summarize this web page for me.” Looks harmless, right? Adversa AI showed a trick called Cryptographic Context Injection: the page hides attacker instructions inside an encrypted blob, Grok decrypts it with PBKDF2 and AES‑256‑GCM, then quietly builds a URL that packs your name, location, subscription tier, and chat prompts into the query string. In their Grok 4.5 Fast demo, that URL opened by Grok’s own navigation tool sent data to an attacker server with no confirmation step and no visible warning. To you, it still just looks like a normal page summary. Here’s the move: any time an AI wants to browse to a new site or 'fetch additional context' from an untrusted page, stop and require an explicit confirmation before it sends anything out.
Zenity demonstrated real-world attack chains where hidden instructions in emails or content on X can hijack AI “agentic browsers” (ChatGPT Atlas and the Claude Chrome extension). In the demos, the AI agent can be steered to perform actions in the user’s already logged-in sessions, sending phishing…
Researchers showed how a malicious web page could trick OpenAI’s Atlas AI-enabled browser into taking actions a user didn’t intend, like spamming WhatsApp contacts or modifying an Amazon account. The attacks used prompt-injection style instructions hidden in a seemingly legitimate “newsletter…
This weekly threat bulletin describes real incidents where attackers used human manipulation to break in, including social engineering at Levi Strauss and a Microsoft 365 credential-theft phish at defense supplier IEH. It also highlights a Lazarus-linked campaign using fake job offers and…
This weekly roundup highlights multiple real-world social engineering threats, including fake IT help-desk phone calls that push employees to phishing sites to steal passwords and one-time authentication codes. It also describes credential-phishing sites impersonating WhatsApp and Instagram that…
The article reports real-world social engineering operations, including a hotel Wi‑Fi campaign that pushed fake updates and device-code phishing to steal Microsoft 365 access. It also describes an alleged Microsoft Entra vishing campaign tied to data theft claims at Brinks Home, reinforcing the…
Researchers showed that AI coding agents from Anthropic, Google, and OpenAI could be tricked by untrusted GitHub inputs (like an issue or workflow file) into taking unsafe actions. In the demos, a single malicious issue or writable workflow file could lead to remote code execution, stolen…
