Reachy Mini robot can be hijacked through five flaws, JFrog security researchers say
Security company JFrog says it found five vulnerabilities in Reachy Mini, the $399-$499 open-source desktop robot from Pollen Robotics and Hugging Face. The robot's software runs a web server on port 8000 with no login, JFrog says, so anything on the same network can tell it to install an app, which runs the attacker's code. The robot's passwordless admin (sudo) access then lets that code take full root control, including its motors, camera and microphone. A second version of the attack works when the owner simply visits a malicious web page. A separate chain pairs a file-upload flaw with two Bluetooth flaws to run code as root from Bluetooth range.
JFrog published a proof-of-concept video and says version 1.11.0 is not vulnerable. Until owners update, it advises blocking port 8000 and not installing untrusted apps. These are JFrog's findings.
More on Hugging Face and home robots.
About the footage: Not used here: JFrog published a proof-of-concept video of the attack on its blog.
The source post
Post by JFrog on X.
View the post on X
Sources
- JFrog: HTTP Vulnerability Allows Hijacking of AI Robot
- JFrog Security Research advisory (CVE-2026-62661)
- @aistasiia on X
New to robots? Read our guide: How robots work: a beginner's guide.
Summarized by Servospace from the sources above, drafted with AI help and checked against those sources before publishing. Claims are the companies' or outlets' unless we say otherwise. How we work.