Loading vulnerability…
MLflow AI Gateway contains a server-side request forgery vulnerability in gateway-secret handling. An authenticated user can supply an unvalidated destination through the gateway secret's auth_config.api_base value. MLflow stores this value and later uses it as the base destination for requests made through the gateway proxy, appending a caller-controlled path and returning the complete response body. The destination is not validated for scheme, host, or IP address range.
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.
1 sources tracked across advisories, community write-ups, and news. New activity surfaces here as Mallory finds it.
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.