CVE-2026-59821 is a post-authentication arbitrary-code-execution vulnerability in LiteLLM Custom Code Guardrails prior to version 1.82.0-stable. The production guardrail create and update flows compiled and executed administrator-supplied Python without applying the sandboxing, built-in restrictions, and forbidden-pattern validation used by the guardrail testing flow. Consequently, a user able to create or modify custom guardrails could execute arbitrary Python in the LiteLLM proxy environment.
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 public exploits tracked yet. Mallory keeps watching.
No public exploit code observed for this vulnerability.
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 LiteLLM Custom Code Guardrails vulnerability allowing an authenticated administrator to execute supplied code, potentially achieving root-level command execution within the LiteLLM container.
A root-level command-execution vulnerability in LiteLLM's Custom Code Guardrails registration endpoint. An authenticated administrator could submit code that executes in the LiteLLM container; retained default master credentials or absent authentication can make administrator access readily available.
A post-authentication custom-code guardrail validation bypass in LiteLLM that allowed Python code execution inside the gateway container, which runs as root in the default Docker image. A proxy-admin credential, including an exposed or default master key, was required.
A post-authentication root-level remote code execution vulnerability in LiteLLM custom code guardrails. The guardrail-registration path executed arbitrary Python without forbidden-pattern validation or removal of Python built-ins. On exposed pre-v1.82.0 instances using the default master key or no authentication, it was effectively pre-authentication RCE.
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.