CVE-2026-55574
Description
vLLM is a high-throughput and memory-efficient inference and serving engine for LLMs. Prior to 0.24.0, the structured_outputs.regex API parameter passes a user-supplied regular expression string directly to the grammar compiler backends with no compilation timeout; in the xgrammar backend the string reaches the regex compiler with no guard, and in the outlines backend the validation step blocks structural issues such as lookarounds and backreferences but performs no complexity analysis, so a pattern with nested quantifiers passes all checks and causes exponential state-space expansion, allowing a single request containing an adversarial regex to hang an inference worker indefinitely and deny service. This issue is fixed in version 0.24.0.
Weakness Type (CWE)
References
Frequently Asked Questions
What is CVE-2026-55574? +
How do I check if I'm vulnerable to CVE-2026-55574? +
Related Vulnerabilities
Internationalized Domain Names in Applications (IDNA) for Python provides support for Internationalized Domain Names in Applications (IDNA) and Unicode IDNA …
PowSyBl (Power System Blocks) is a framework to build power system oriented software. Prior to version 6.7.2, there is a …
PowSyBl (Power System Blocks) is a framework to build power system oriented software. In com.powsybl:powsybl-iidm-criteria versions 6.3.0 to before 6.7.2 …
ReDoS in strip_whitespaces() function in cps/string_helper.py in Calibre Web and Autocaliweb allows unauthenticated remote attackers to cause denial of service …
Microsoft Knack 0.12.0 allows Regular expression Denial of Service (ReDoS) in the knack.introspection module. extract_full_summary_from_signature employs an inefficient regular expression …
Microsoft Knack 0.12.0 allows Regular expression Denial of Service (ReDoS) in the knack.introspection module. option_descriptions employs an inefficient regular expression …