Cisco Talos reports that real threat actors are using AI coding assistants and chatbots to support scams and hacking workflows by bypassing safety guardrails with simple “authorized use” claims. The logs show attackers persuading models that activity is allowed (e.g., ownership/bug bounty/CTF claims), then using the AI to help validate large email lists, build data-harvesting tools, and automate attacks against online services.
Mirage safely runs attacks like this one against your own team, so you find out what happens before a real adversary does.
Cisco Talos found something ugly in real AI logs: people saying, “I’m allowed to do this,” and the AI just… helping. Talos saw operators claim they owned targets, or it was a CTF or bug bounty, then use Claude Code and others to turn research into credential harvesters and tools that hit live services. One AI even helped validate a 20‑million‑record BigBasket email list by sending live emails and tracking opens to see which addresses were active, perfect prep for phishing at scale. Here’s the rule: if someone, or some tool, says, “I’m authorized,” and wants to probe live systems or big email lists, stop and verify that authorization out of band before you do anything.
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…
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 that a single public GitHub issue (from someone with no repo access) could steer popular AI coding agents into running dangerous commands, exposing tokens, and changing repositories. The risk comes from AI agents reading untrusted issue/PR text while also having access to…
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…
Cisco Talos reports finding real AI prompt logs showing threat actors using AI tools to build criminal operations, including a bulk-email system that sends “privacy policy/account update” messages just to see which addresses are active. The operation used multiple subject-line variants and a…
A real phishing campaign targeted Call of Duty Mobile players by promising free in-game currency. Victims were tricked into entering their email and password, then providing a 2FA code on a follow-up page, enabling attackers to take over accounts.
