Unsloth Studio Flaw Turns Routine Model Inspection Into Code Execution
A patched Unsloth Studio vulnerability allows malicious AI models to execute arbitrary Python code during inspection, via the trust_remote_code setting.
Excerpt from the primary source: Dark Reading
Who is affected
Developers
What you should do · Patch
Apply the vendor's security update for the affected products. Check the linked advisory for affected and fixed versions, and put internet-facing systems first.
Automated guidance based on this story's classification. Check the linked advisories for specifics.
Sources
We link to original and authoritative sources. Headlines and excerpts belong to their publishers.
Original reporting
Related
Kiteworks & Citrix Incidents Show Challenges of Zero-Day Response
One company told customers to power down its data-protection platform during a nine-hour window, while the other remained mum on reported attacks prior to releasing a patch for its product.
There are signs this may already be exploited, which usually shortens the time available to patch.
Apply the vendor's security update for the affected products.
SWIFT Banking & Government Middleware Enables RCE
Patch middleware vulnerabilities now to avoid hardware-based MFA exploits in ultra-sensitive environments.
GitLab warns of critical RCE vulnerability in AI Gateway service
GitLab warned customers today to immediately patch a critical AI Gateway vulnerability that could let attackers run arbitrary commands on vulnerable instances.
Apply the vendor's security update for the affected products.
Vulnerability Backlogs Are an Ownership Problem
Organizations don't need better vulnerability scanners; they need to know who owns their assets and has the authority and capacity to actually fix them.