Track vulnerabilities like this one.
Sign up free to get alerted when software you use is affected.
TensorFlow 2.16.1-11 may unintentionally send data to a rogue TFTP server
published 15 days ago
Summary
TensorFlow version 2.16.1-11 uses libcurl, which can be tricked by a malicious TFTP server into switching from download mode to upload mode. This can cause the program to transmit data that was meant for other operations, potentially exposing sensitive information. Update TensorFlow to a version that includes the libcurl fix or restrict TensorFlow from contacting untrusted TFTP servers.
What to do
The CVE record does not list a fixed version. Check the vendor's site or the advisory links below - a fix may already be released.
Affected software
| Ecosystem | Vendor | Product | Affected versions |
|---|---|---|---|
| Azure Linux:3 | azure | tensorflow | <= 2.16.1-11 |
Original advisory text
CVE-2026-92389 affecting package tensorflow 2.16.1-11
libcurl's TFTP state machine fails to validate state transitions against the
requested transfer direction, accepting opening server responses for both
reads and writes interchangeably. A malicious TFTP server can respond to a
download request in a way that forces the client into transmit mode, prompting
libcurl to consume and transmit data from the handle's persistent
`CURLOPT_READDATA` pointer - including data remaining from unrelated prior
operations, such as an HTTP POST request executed on a reused handle.
requested transfer direction, accepting opening server responses for both
reads and writes interchangeably. A malicious TFTP server can respond to a
download request in a way that forces the client into transmit mode, prompting
libcurl to consume and transmit data from the handle's persistent
`CURLOPT_READDATA` pointer - including data remaining from unrelated prior
operations, such as an HTTP POST request executed on a reused handle.
References
Timeline
Published20 Oct 2026
Updated23 Oct 2026
First seen23 Oct 2026
Sources
AZL-110301 · OSV
Track software like this
Free during beta