Machine-written summary
Summary written by Claude Haiku from the source advisory. Not reviewed by a person.
xgrammar versions before 0.1.18 keep an unbounded in-memory cache of compiled grammars. An attacker who can send requests to a system using xgrammar, such as an LLM inference server, with many distinct JSON schemas can fill the host's memory and cause a denial of service. Upgrade xgrammar to 0.1.18 or later, which adds a limit to the cache size.
Not analysed yet
We haven’t published our analysis of this Advisory yet. Once a person has reviewed it, this page will say what’s vulnerable, what has to be true for your app to be affected, and how to check.
Severity and exploitation
| Score | Rating | CVSS | Vector | Source |
|---|---|---|---|---|
| 6.5 | Medium | 3.1 | CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H | CNA (GitHub_M) |
- CISA KEV Not listed.
- EPSS 0.0052, the 44th percentile, on 17 Nov 2026: FIRST’s estimate of the chance of exploitation in the next 30 days.
Timeline
- 31 Oct 2026, 14:29 UTC First seen by agentprime
- 31 Oct 2026, 14:29 UTC Confirmed by OSV
Sources
- GitHub Advisory Database: GHSA-389x-67px-mjg3
- OSV: GHSA-389x-67px-mjg3
- github.com/mlc-ai/xgrammar/security/advisories/GHSA-389x-67px-mjg3
- nvd.nist.gov/vuln/detail/CVE-2025-32381
- github.com/mlc-ai/xgrammar/pull/243
- github.com/vllm-project/vllm/pull/16283
- github.com/mlc-ai/xgrammar
- github.com/pypa/advisory-database/tree/main/vulns/xgrammar/PYSEC-2025-235.yaml