Process abort via merge buffer overrun
Published Sep 4, 2026 · Updated Sep 4, 2026
Out-of-bounds write in BpeBuilder in Hugging Face Tokenizers 0.23.2 and earlier allows attackers to abort a process via crafted tokenizer.json files. BpeBuilder::build sizes a scratch buffer to the longest vocabulary key, then copies each concatenated merge token without ensuring that it fits. Loading an attacker-supplied file through Tokenizer::from_file or from_str reaches the write before validation, causing a panic that can terminate Rust and FFI-hosted services.
Summary
What happened
Out-of-bounds write in BpeBuilder in Hugging Face Tokenizers 0.23.2 and earlier allows attackers to abort a process via crafted tokenizer.json files. BpeBuilder::build sizes a scratch buffer to the longest vocabulary key, then copies each concatenated merge token without ensuring that it fits. Loading an attacker-supplied file through Tokenizer::from_file or from_str reaches the write before validation, causing a panic that can terminate Rust and FFI-hosted services.
The record
- CVE
- CVE-2026-85670
- Published
- Sep 4, 2026
- Updated
- Sep 4, 2026
- Vendor
- Hugging Face
- Product
- Tokenizers
- Classifications
- CWE-787, T1499.004
- Attack vector
- network
- Privileges
- unauthenticated
Timeline
How it unfolded
- Sep 4, 2026CVE publishedPublication date reported by the CVE source.
- Sep 4, 2026Record 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 <=0.23.2
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.
Public exploit references
- BPE merge buffer-overrun 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