Sunday , September 27 2026
LiteLLM

LiteLLM supply chain attack reveals 153GB of stolen credentials online

153GB record surface online stolen during the LiteLLM supply chain attack linked to thousands of corporate domains, including AWS, Samsung, Cisco, and Salesforce. Hudson Rock said the archive has 433,909 files, and linked 118,829 CI runner dumps to 2,488 company domains.

“We are leveraging this data for a global ethical disclosure effort,” Alon Gal, Hudson Rock’s co-founder and CTO, told Help Net Security. “We see this as an opportunity to help organizations respond proactively before threat actors weaponize the data publicly.”

Microsoft Patches CVSS 10.0 Azure AI Foundry Vulnerability Allowing Privilege Escalation

Microsoft has fixed a serious security flaw in Azure AI Foundry that could let bad actors gain privilege escalation. The...
Read More
Microsoft Patches CVSS 10.0 Azure AI Foundry Vulnerability Allowing Privilege Escalation

AWS is unable to restore access to Bahrain, one UAE cloud data zone after war damage

Amazon Web Services cannot restore access to its cloud-computing facility in Bahrain and ‌one of three data-hosting zones in the...
Read More
AWS is unable to restore access to Bahrain, one UAE cloud data zone after war damage

Cisco Warns of Critical ISE 0-Day Flaw and Hackers Allegedly Selling Fortinet FortiGate 1-Day Flaw

A threat actor is allegedly offering a private remote code execution exploit for Fortinet FortiGate SSL VPN appliances, claiming that...
Read More
Cisco Warns of Critical ISE 0-Day Flaw and Hackers Allegedly Selling Fortinet FortiGate 1-Day Flaw

Anthropic prepares “Claude Money” to analyze bank account and financial data

Anthropic is making a new Claude feature called “Money.” It's a separate tab in the mobile app. The new interface...
Read More
Anthropic prepares “Claude Money” to analyze bank account and financial data

GhostCode Phishing Kit Evades Microsoft 365 MFA to Hijack Accounts in 78 Seconds

GhostCode is a new phishing kit that changes a regular Microsoft 365 sign-in into an account theft. It doesn't need...
Read More
GhostCode Phishing Kit Evades Microsoft 365 MFA to Hijack Accounts in 78 Seconds

CISA Warns of Cisco Secure Email Gateway 0-Day Flaw Actively Exploited in Attacks

CISA has added a serious Cisco Secure Email Gateway flaw to its list of known exploits. They warn that attackers...
Read More
CISA Warns of Cisco Secure Email Gateway 0-Day Flaw Actively Exploited in Attacks

VPN flaw exposed 246,000 personnel records in japan

Japan’s Digital Agency found a data leak that may have exposed about 246,000 records with personal information of government workers....
Read More
VPN flaw exposed 246,000 personnel records in japan

Hackers deploy Casbaneiro Trojan that activates on bank websites

Casbaneiro is going after online banking users by sending fake messages that seem like urgent bills or legal papers. The...
Read More
Hackers deploy Casbaneiro Trojan that activates on bank websites

German police read Signal, Telegram, WhatsApp messages without breaking encryption

German law enforcement agencies are using features built into apps such as WhatsApp to monitor people’s messages without breaking their...
Read More
German police read Signal, Telegram, WhatsApp messages without breaking encryption

Urgent Patch! cPanel, GitLab Flaws Expose Users to RCE, File and Credential Theft

GitLab has released an important security update to fix two serious problems. These issues could allow unauthorized file access and...
Read More
Urgent Patch! cPanel, GitLab Flaws Expose Users to RCE, File and Credential Theft

On March 19, 2026, TeamPCP, a group of cybercriminals that started in late 2025, used stolen login info to release a hacked version of Trivy. LiteLLM’s build system installed Trivy on its own, giving the bad scanner access to the runner area and letting the attackers take the project’s PyPI publishing keys. TeamPCP released two harmful LiteLLM versions, 1.82.7 and 1.82.8, to the Python Package Index on March 24.

“The key takeaway is how supply chains have evolved to make a single upstream breach affect thousands of companies simultaneously. A window of roughly 40 minutes in which the LiteLLM dependency was hacked led to over 430,000 instances in which millions of secrets were harvested,” Gal said.

Tracing exposed secrets back to their owners

The data includes details about companies like NVIDIA, Volkswagen, Microsoft, FedEx, S&P Global, John Deere, Epic Games, Orange, TomTom, BT Group, ServiceNow, Deloitte, and Siemens.

Screenshots from the research show secret access keys for AWS, client secrets for Salesforce, signing secrets for Slack, environment variables for Azure, and API keys for AI providers taken during pipeline runs.

A trove of internal corporate secrets (Source: Hudson Rock)

CloudSEK estimated from a different dataset that nearly 500 organizations are exposed.

“I’ve confirmed the data is legit by the way, multiple victim orgs,” security researcher Kevin Beaumont said. “It contains a significant volume of sensitive content at orgs. It’s a massive supply chain breach due to poor AI security—not because AI is the threat, but teens can run circles around orgs obsessed with rushing out AI and poor DevOps security.”

“Identifying the victims within this 153GB database presents a unique threat intelligence challenge. Accurate attribution requires looking past the surface level to analyze the actual infrastructure boundaries,” Hudson Rock noted.

A leaked pipeline was connected to a committer email at SiriusXM. Other signs in the same leak, like a self-hosted GitLab found at gitlab.adswizz.com, led to AdsWizz, which is a part of SiriusXM.

Hudson Rock said correct identification depends on “hard infrastructure markers rather than simple committer emails.”

Many of the dumped files do not have a clear owner. Hudson Rock says some have database passwords, third-party API keys, and cloud login details without company emails, custom domains, or server names that show which organization they belong to.

Organizations urged to act before stolen data leaks

Organizations urged to act before stolen data leaks Hudson Rock is asking organizations that use AI proxy tools, third-party security scanners, or AI packages to check their systems for LiteLLM versions 1.82.7 and 1.82.8. They should treat any secrets in the LiteLLM environment as unsafe.

Organizations should change cloud IAM keys and access tokens often. They need to look at audit logs for strange activity since March 24 and find any unauthorized .pth files and suspicious systemd services.

“While we cannot disclose how we obtained this data, it is not leaked anywhere at the moment and is not circulating widely. This makes it a critical window of opportunity for companies to rotate keys and secrets before it eventually leaks, as is the natural cycle of a breach,” Gal added.

Many groups seem to be dealing with the breach less seriously than they should, even though it is significant.

“These creds date from about March. One of the orgs impacted told me they’d rotated them all and it’s a nothingburger, so I looked at their responsible disclosure policy, it allows trying creds, so I tried them all. Almost every one worked. Submitted report. One of the biggest US techcos,” Beaumont wrote.

“This magnitude pushes us into a completely new world regarding the type of response required from the cybersecurity industry,” Gal concluded.

Check Also

AI models

CISA Says Chinese Firms Extracted Billions of Tokens From Frontier AI Models

Six Chinese AI companies ran large-scale attacks on American AI models since late 2024, according …