## vLLM: DoS via unbounded chat_template Jinja2 rendering exhausts the renderer thread pool

- 種別: advisory
- 重大度: medium
- 関連技術: python
- CVE: CVE-2026-115911
- 公開日時: 2026-10-28T21:14:52.000Z
- 情報源: GitHub Advisory Database (https://github.com/advisories/GHSA-542m-9234-8gvf)
- パーマリンク: https://vulnpulse.app/items/ghsa-ghsa-542m-9234-8gvf

影響: pip:vllm >= 0.19.0, < 0.32.0 → 0.32.0
vLLM renders chat templates by passing the resolved Jinja2 source to tokenizer.apply_chat_template(), which compiles and .render()s it in an ImmutableSandboxedEnvironment. The sandbox prevents code-execution escapes but imposes no wall-clock timeout, no output-size cap, and no nested-loop or binop limit. The render runs in a ThreadPoolExecutor whose max_workers defaults to 1, and CPython worker threads cannot be interrupted. An authenticated API client (or a malicious model repo loaded with tru…