This blog details research and analysis of an active campaign that exploits a critical unauthenticated remote code execution (RCE) vulnerability, CVE-2025-3248, that has been identified in Langflow versions prior to 1.3.0. Langflow is a Python-powered visual framework for building AI applications with over 70,000 GitHub stars, and its versions prior to 1.3.0 contains a flaw in its code validation mechanism that permits arbitrary code execution.
Unauthenticated attackers can exploit this vulnerability by crafting malicious POST requests to the /api/v1/validate/code endpoint. The malicious payload in our investigation was found embedded within argument defaults or decorators of a Python function definition. Since Langflow does not enforce input validation or sandboxing, these payloads are compiled and executed within the server’s context, leading to RCE.
Read more…
Source: Trend Micro
Sign up for our Newsletter
The latest news and insights delivered right to your inbox.
Related:
- Experts working with Homeland Security hacked into Boeing 757
November 10, 2017
There’s some unsettling news about one of America’s most widely-used jetliners. In a test, experts working with Homeland Security hacked into a Boeing 757. The team of researchers needed only two days in September 2016 to remotely hack into a 757 parked at the airport in Atlantic City, New Jersey. Speaking at a conference this week, Robert Hickey of ...
- Intel’s management engine – in most CPUs since 2008 – can be p0wned over USB
November 9, 2017
Positive Technologies, which in September said it has a way to attack the Intel Management Engine, has dropped more details on how its exploit works. The firm has already promised to demonstrate God-mode hack in December 2017, saying the bug “allows an attacker of the machine to run unsigned code in the Platform Controller Hub on any motherboard”. For ...
- Evil pixels: researcher demos data-theft over screen-share protocols
November 9, 2017
It’s the kind of thinking you expect from someone who lives in a volcano lair: exfiltrating data from remote screen pixel values. The idea comes from Pen Test Partners’ Alan Monie, taking a break from sex toy hacks and wondering how to get data over a connection like RDP (remote desktop protocol) when the target had blocked file transfer ...
- Russian ‘Fancy Bear’ Hackers Using (Unpatched) Microsoft Office DDE Exploit
November 8, 2017
Cybercriminals, including state-sponsored hackers, have started actively exploiting a newly discovered Microsoft Office vulnerability that Microsoft does not consider as a security issue and has already denied to patch it. Last month, we reported how hackers could leverage a built-in feature of Microsoft Office feature, called Dynamic Data Exchange (DDE), to perform code execution on the targeted device ...
- Hacker Distributes Backdoored IoT Vulnerability Scanning Script to Hack Script Kiddies
November 8, 2017
Nothing is free in this world. If you are searching for free hacking tools on the Internet, then beware—most freely available tools, claiming to be the swiss army knife for hackers, are nothing but a scam. For example, Cobian RAT and a Facebook hacking tool that we previously reported on The Hacker News actually could hack, but of the one who ...
- US-CERT Warns of Crypto Bugs in IEEE Standard
November 6, 2017
Recent academic work focused on weak cryptographic protections in the implementation of the IEEE P1735 standard has been escalated to an alert published Friday by the Department of Homeland Security. DHS’ US-CERT warned the IEEE P1735 standard for encrypting electronic-design intellectual property and the management of access rights for such IP is flawed. “In the most egregious cases, enable attack vectors that allow ...