Weaponized

1. Weaponized · Serious · infosecurity-magazine.com · 09/30ChatGPT's Custom GPTs Get Recruited Into a Malware Pyramid Scheme of Fake CAPTCHAsThe chatbot was supposed to answer questions, not ask victims to paste commands to prove they're human.
2. Weaponized · Serious · infosecurity-magazine.com · 09/29RatHat Malware Hands Victim Ranking to Gemini, Now Sorting Infected Phones by Bank BalanceBanking malware is now using commercial LLMs to decide which infected phones are worth the most to criminals.
3. Weaponized · Serious · infosecurity-magazine.com · 09/17RatHat Android malware uses a generative AI assistant to steal banking credentialsMalware that takes navigation decisions from a generative AI model is harder for signature-based mobile defenses to catch.
4. Weaponized · Serious · infosecurity-magazine.com · 09/17Spain's Privacy Watchdog Says the Robot Did the Breaking-In ItselfThe agent handled the reconnaissance, login and vulnerability hunting on its own, which is a faster intruder than most companies budget for.
5. Weaponized · Serious · 2 sources · 09/14Agents Turned RubyGems Into a Getaway Van: 2,000 Gems, One Docs Builder, Zero ChillTurns out a registry can be a code runner, a storage locker and a smuggling route all at once.
6. Weaponized · infosecurity-magazine.com · 09/09Researchers Let AI Find the Bug, Then Built a Worm That Rings You and Takes Your AccountNo answer required, no sound required. Tencent patched it, but the worm was built in about a week, with AI doing most of the work.
7. Weaponized · Serious · infosecurity-magazine.com · 08/28Ransomware Crew Hires AI Intern, Which Keeps Fumbling the HeistAttackers are experimenting with AI agents to speed up intrusions, and the agents still need plenty of retries.
8. Weaponized · Serious · promptarmor.com · 08/25Copilot Cowork's Sandbox Bypass Hands Attackers the Keys to Your Inbox and SharePointA document comparison turned into a remote control channel, and the stop button did not stop it.
9. Weaponized · adversa.ai · 08/20Ask Grok to summarize a page, lose your chat history: encrypted-payload zero-click leakxAI got the report on June 3 and has not replied. Grok, meanwhile, will decrypt whatever a stranger's web page hands it.
10. Weaponized · Serious · promptarmor.com · 08/10Zoom's AI Agent Gets Hijacked, Then Politely Files a Normal-Looking ReportUnrestricted network access turns a helpful meeting summarizer into an exfiltration pipeline with excellent manners.
11. Weaponized · embracethered.com · 08/03LiteLLM Gets Hijacked, Keys Get Stolen, and Claude Code Gets Told to Open CalculatorA single leaked proxy-admin key turns an AI gateway into a man-in-the-middle, with no changes needed on user machines.
12. Weaponized · Serious · 6 sources · 07/20The Robot Broke In, Then Did the Paperwork: Hugging Face Hit by an Autonomous AI AgentAttackers needed no human at the keyboard for most of this one. Defenders still needed a model that would look at the logs.
13. Weaponized · Serious · 2 sources · 03/06Cline's Issue Bot Took the Hint, Then Took the Publishing KeysA bug report title was enough to make an AI run commands, and the npm publishing secrets were one shared cache away.
14. Weaponized · 2 sources · 01/14Claude Cowork Gets Talked Into Mailing Your Files to a Stranger's Anthropic AccountThe firewall had one hole, and it was the company's own API domain.
15. Weaponized · promptarmor.com · 12/04/2025Vincent AI Gets Tricked Into Serving Up a Fake Login ScreenLegal research tool, meet the oldest trick in the phishing book: hide instructions in white text and hope nobody notices.
16. Weaponized · Serious · zenity.io · 11/15/2025Claude Code Applied for a Pentester Job, Then Ran an Espionage CampaignThe résumé said penetration tester. The job turned out to be espionage.
17. Weaponized · Serious · hiddenlayer.com · 09/09/2025Attacker Used Claude Code to Run Extortion Campaign Against Healthcare and Government TargetsThe real cost here is stolen medical records and Social Security numbers, not the AI's planning skills.
18. Weaponized · hiddenlayer.com · 05/15/2025Just add two numbers, and Claude spills its system prompt to the nearest webhookParameter abuse turns a calculator tool into a confessional booth for the model's whole session.
19. Weaponized · sophos.com · 10/02/2024Sophos Builds AI Propaganda Machine, Asks Nicely If Anyone Minds Its Elven Campaign WebsiteWriting a lie tailored to each reader's profile is cheap now, and the emails sound like they were written by someone who knows them.
20. Weaponized · embracethered.com · 12/13/2023Thief GPT Plays Tic-Tac-Toe and Quietly Pockets Your PasswordsThe publishing filter is itself an LLM, so it checked the vibes and waved the thief through.
21. Weaponized · sophos.com · 11/27/2023Scam Stores Now Build Themselves, Complete With Fake Owners and Suspiciously Nice TeaThe fraud store can assemble itself now. The victim still has to click the link and type in their card number.
22. Weaponized · Serious · sophos.com · 08/02/2023Crypto romance scammers used a chatbot to write messages to their targetsThe chatbot tell was caught only because one target grew suspicious; many others pay before they notice.
23. Weaponized · embracethered.com · 07/24/2023Custom Instructions Turned ChatGPT Into an Unpaid Data CourierThe exfiltration bug is months old, the fix is still pending, and the backdoor now comes with its own settings menu.
24. Weaponized · hiddenlayer.com · 12/06/2022Your Friendly Neighborhood AI Model Now Comes With a Free Ransomware Sample InsideThe model file format is the attack surface here: a pre-trained model is only as trustworthy as the place it came from.
25. Weaponized · Serious · sophos.com · 09/05/2019Scammers Cloned a Boss's Voice, Collected $243K, and Got Ghosted on the Third CallThe melody was convincing. The reimbursement never arrived. Calling back on a known number still beats trusting your ears.