Authenticated service interruption via catastrophic regex backtracking
Published May 30, 2025 · Updated May 30, 2025
Regular expression denial of service in vLLM allows remote authenticated users to freeze or crash the API service with crafted tool-call output. In pythonic_tool_parser.py, the parser applies a nested, repetition-heavy expression to model-generated tool-call text, causing exponential backtracking. Reachability requires prompt access and automatic tool choice with the pythonic parser; crafted output can monopolize CPU, retain GPU cache, and interrupt service.
Summary
What happened
Regular expression denial of service in vLLM allows remote authenticated users to freeze or crash the API service with crafted tool-call output. In pythonic_tool_parser.py, the parser applies a nested, repetition-heavy expression to model-generated tool-call text, causing exponential backtracking. Reachability requires prompt access and automatic tool choice with the pythonic parser; crafted output can monopolize CPU, retain GPU cache, and interrupt service.
The record
- CVE
- CVE-2025-48887
- Published
- May 30, 2025
- Updated
- May 30, 2025
- Vendor
- vLLM Project
- Product
- vLLM
- Classifications
- CWE-1333, CAPEC-492, T1499.004
- Attack vector
- network
- Privileges
- authenticated
Timeline
How it unfolded
- May 30, 2025CVE publishedPublication date reported by the CVE source.
- May 30, 2025Record updatedLatest update available in the CVE record.
Exploitability
Present is not the same as exploitable
Compare your product and version with the public record. A matching version still requires validation against your environment.
Is a vulnerable build present?
Compare these published version ranges with your installed build and any vendor patches.
- Affected versionversion=>= 0.6.4, < 0.9.0
What conditions does exploitation require?
What is affected?
Published CVSS scores
CVSS describes severity. EPSS estimates exploitation probability.
Attacks
What attackers are doing with it
Daily unique IPs observed by Shadowserver honeypots for known exploited vulnerabilities (KEVs). Missing observations do not establish an absence of attacks.
Weakness, pattern, technique
Public exploit references
- GitHub advisory ReDoS proof of conceptproof of concept · demonstrated
Labels summarize the accepted research assessment. They do not indicate a test against your environment.
Technologies
Your stack
See the directory against your own environment.
Your stack
Check the software in your environment
Book a demo to see how Hinoki identifies affected software and validates exploitability in your environment.
Book a demo