Loading vulnerability…
CVE-2024-0520 is a path traversal vulnerability in MLflow prior to version 2.9.0 affecting HTTP dataset loading in the mlflow.data.http_dataset_source.py module. When MLflow downloads a dataset from an HTTP source, it derives the output filename from the HTTP Content-Disposition header or source URL path without adequately sanitizing path components. An attacker controlling the HTTP response or source URL can supply relative traversal sequences or an absolute pathname, causing MLflow to write the downloaded content outside the intended dataset-download directory. The issue encompasses both relative and absolute path traversal and can result in arbitrary file overwrite.
Mallory correlates every CVE against your assets, your vendors, and active adversary campaigns. Know which vulnerabilities matter for you, not just which ones are loud.
What it means. What to do now. Patch path, mitigations, and the assume-compromise checklist.
What an attacker gets, and what they’ve been doing with it.
If you can’t patch tonight, do this now.
Patch, then assume compromise.
No valid public exploits. Mallory filtered out 1 candidate as fakes, detection scripts, or README-only repos.
All candidate exploits were filtered out by Mallory's validation.
Products and vendors Mallory has correlated with this vulnerability. Open in Mallory to drill down to specific CPE configurations and version ranges.
Vendor-confirmed product mapping. Mallory continuously reconciles this list against your asset inventory.
7 sources tracked across advisories, community write-ups, and news. New activity surfaces here as Mallory finds it.
A path traversal vulnerability in MLflow described in the content as enabling remote code execution via OS command injection.
A path traversal vulnerability in a dataset management product for AI model training and evaluation that allows file overwrite via the Content-Disposition header.
Path traversal vulnerability in a product for managing datasets for AI model training and evaluation, allowing file overwrite via the Content-Disposition header.
Path traversal allowing file overwrite through a Content-Disposition header in an AI-training/evaluation dataset-management product.
Query your assets running an affected version, and investigate the blast radius.
Every observed campaign linking this CVE to a named adversary.
Malware families riding this exploit, with evidence and IOCs.
YARA, Sigma, Snort, and vendor rules, auto-deployed to your SIEM.
Cross-references every affected SKU, including bundled OEM variants.
Community discussion across Reddit, Mastodon, and other social sources.