RCE via Hydra instantiate() on untrusted model metadata in Salesforce Uni2TS (<= 1.2.0)
Salesforce Uni2TS through 1.2.0 is vulnerable to code injection leading to remote code execution when loading a crafted/poisoned model that supplies attacker-controlled configuration/metadata. The issue arises from Uni2TS passing untrusted model configuration data (e.g., from Hugging Face model config/metadata) into Meta Hydra’s hydra.utils.instantiate() without sufficient validation/allowlisting. Because Hydra instantiate() can invoke arbitrary Python callables specified via configuration (e.g., builtins.exec/eval or os.system) and attacker-controlled arguments, a malicious model can trigger execution of attacker-supplied code during model load/initialization on macOS, Windows, or Linux.
Are you exposed to this one?
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.
Impact, mitigation & remediation
What it means. What to do now. Patch path, mitigations, and the assume-compromise checklist.
Impact
What an attacker gets, and what they’ve been doing with it.
Mitigation
If you can’t patch tonight, do this now.
Remediation
Patch, then assume compromise.
Exploits
No public exploits tracked yet. Mallory keeps watching.
No public exploit code observed for this vulnerability.
Affected products & vendors
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.
Recent activity
11 sources tracked across advisories, community write-ups, and news. New activity surfaces here as Mallory finds it.
An RCE vulnerability affecting Salesforce’s Uni2TS AI/ML Python library where malicious model metadata can trigger arbitrary code execution through Hydra’s instantiate behavior.
Remote code execution risk in Salesforce’s Uni2TS where configuration/metadata handling results in untrusted data reaching Hydra’s instantiate(), allowing invocation of arbitrary callables during model/config loading.
Remote code execution risk in Salesforce’s Uni2TS Python AI library caused by passing untrusted model metadata into Hydra’s instantiate() without sufficient validation, enabling execution of attacker-chosen callables during model load.
Remote code execution risk in Salesforce’s Uni2TS Python AI library when loading a compromised model file: attacker-controlled metadata is used with Hydra’s instantiate() without proper input validation, enabling execution of arbitrary callables.
The version that knows your environment.
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.