CVSS:3.1/AV:N/AC:L/PR:H/UI:N/S:U/C:H/I:H/A:H
Called exploited 18 days after disclosure.
Measured from the CVE publication date to the earliest of 1 KEV catalogue that list it.
The life of this vulnerability
- CVE published
- First KEV listing18d
- Last sightingsame day
Gaps are compressed to equal steps. The elapsed time is printed under each.
Which catalogues call it exploited
- CISA KEVUS federaldoes not list it
- EUVDENISA, European Uniondoes not list it
- VulnCheck KEVcommercial researchlisted Jul 26, 2026
- CIRCLaggregator, mirrors the abovedoes not list it
This rests on a single catalogue. No second catalogue corroborates the claim that it is being exploited. CIRCL is an aggregator and is not counted.
Public exploitation evidence
- reported exploitationwww.linkedin.com/posts/yaarashriki_cve-2026-59821-cve-2026-5
1 public report collected from VulnCheck and CIRCL, first on Jul 26, 2026. Each links to its original source. We have not verified them.
Description
LiteLLM is a proxy server (AI Gateway) to call LLM APIs in OpenAI (or native) format. Prior to 1.82.0-stable, LiteLLM's Custom Code Guardrails production create and update paths did not apply the same sandboxing and validation used by the test endpoint, allowing a privileged user with access to create or update guardrails to submit custom Python code that executed in the LiteLLM proxy environment and could expose secrets available to the process. This issue is fixed in version 1.82.0-stable.