SUPPLY // AI threat desk

AI Supply Chain_

Malicious models, poisoned packages, plugins and datasets

10 stories

AI-generated image of a security engineer reviewing server dependency files on blurred monitors in an operations centre.

Corelight: MCP server setups let dependencies change unnoticed

Corelight said 23 of 35 reviewed MCP servers had recommended installs that would resolve dependencies fresh from a public registry. It published the findings on Nov 10, 2026.

14 Nov 20263 min read

AI-generated image of security analysts reviewing blurred cloud access logs in a dim operations centre.

SUPPLYAWS

AWSMedium

Qubole AWS integrations abused to access customer environments

Wiz observed a threat actor using IAM credentials from an AWS account associated with Qubole between Oct 14 and Oct 26, 2026. Wiz published its report on Nov 9.

10 Nov 20263 min read

AI-generated image of engineers monitoring cryptocurrency exchange systems beside blurred dashboards.

SUPPLYAISEC

AISECMedium

Binance said campaign culminated in $6.43 billion in assets

Binance said a multi-stage campaign culminated in $6.43 billion in assets being taken from the cryptocurrency exchange, involving two engineers and an internal AI tool.

4 Nov 20262 min read

AI-generated image of security analysts monitoring a cyberattack response in a dim operations centre.

SUPPLYXCPE

XCPEMedium

Google disrupted XCPE botnet and banned Videostrong after 1.3M infections

Google and partners dismantled XCPE after it enrolled more than 1.3 million Android devices and banned Videostrong from certified Android work. Google Cloud disclosed it Nov 2.

3 Nov 20263 min read

AI-generated image of security analysts examining a suspicious software package alert in a dim operations centre.

SUPPLYPYPI

PYPIMedium

LiteLLM 1.107.0 malicious release could expose credentials and AI tokens

Cybersecurity News said a malicious LiteLLM 1.107.0 release reportedly reached PyPI with a credential stealer and remote-access malware. HKCERT disclosed the findings on Oct 31, 2026.

2 Nov 20262 min read

AI-generated image of a security analyst reviewing blurred package and container security dashboards in an operations centre.

SUPPLYSUPPLY

SUPPLYHigh

LiteLLM pipeline compromise published compromised version 1.107.0

LiteLLM published compromised 1.107.0 packages to PyPI and container registries on Oct 21. LiteLLM disclosed the incident on Oct 23.

26 Oct 20262 min read

AI-generated image of security analysts examining a compromised website plugin in a monitoring centre.

SUPPLYCVE

CVECritical

Rank Math PRO malware could compromise WordPress sites

Wordfence classified the event as a supply-chain compromise and rated it 9.8 on the CVSS scale; it is tracked as CVE-2026-111862.

23 Oct 20262 min read

AI-generated image of a cybersecurity engineer reviewing a blurred container registry account page beside server racks.

SUPPLYDOCKER

DOCKERCritical

Docker Hub flaw could let attackers impersonate users and publish images

Docker disclosed a critical Docker Hub flaw that could let attackers impersonate users and publish images under trusted names. The flaw affected personal access tokens transferred during account conversions.

14 Oct 20263 min read

SUPPLYRCE

RCEHigh

Unsloth Studio flaw could run code from a malicious model

Unsloth Studio could execute Python code from a selected model repository, exposing data and credentials accessible to the backend. Pillar Security disclosed the flaw on Sep 29, 2026.

2 Oct 20263 min read

Illustration of an autonomous robot patching a vulnerability between a secure messaging phone and server rack.

SUPPLYAISEC

AISECMedium

DARPA selects Xint to research autonomous AI security for messaging apps

DARPA selected Xint to research autonomous AI security for messaging applications used throughout the Department of War, SecurityWeek reported.

1 Oct 20262 min read