#

CVE-2026-126301: Quadratic CPU Exhaustion in NLTK LegalitySyllableTokenizer

Vulnerability ID: GHSA-GPWC-27CW-RH9R CVSS Score: 5.3

Published: 2026-11-13 The Natural Language Toolkit (NLTK) versions 3.6 through 3.10.3 contain an algorithmic complexity vulnerability in the LegalitySyllableTokenizer.tokenize method. When processing tokens composed of continuous runs of vowels, an unguarded string reversal and set membership lookup loop executes in quadratic time relative to the input token length. Remote attackers capable of submitting arbitrary text to applications utilizing this tokenizer can trigger sustained single-core CPU exhaustion and application-level denial of service.

#

TL;DR

An unguarded loop in NLTK's LegalitySyllableTokenizer.tokenize method causes quadratic O(n^2) CPU exhaustion when processing long vowel sequences, allowing unauthenticated attackers to cause a denial of service.

⚠️ Exploit Status: POC

#

Technical Details

CVE ID : CVE-2026-126301 (GHSA-gpwc-27cw-rh9r) #

Weakness Class : CWE-407 (Inefficient Algorithmic Complexity) / CWE-400 #

Affected Component : nltk.tokenize.legality_principle.LegalitySyllableTokenizer.tokenize #

CVSS v3.1 / v4.0 : 5.3 Medium (v3.1) / 6.9 Medium (v4.0) #

EPSS Score : 0.51% (44.01st Percentile) #

Impact : Single-core CPU Exhaustion / Application Denial of Service #

Exploit Status : Public Proof-of-Concept (PoC) Available #

Patched Version : nltk 3.10.4

#

Affected Systems

• Python applications and NLP pipelines utilizing nltk.tokenize.LegalitySyllableTokenizer

nltk : >= 3.6, < 3.10.4 (Fixed in:3.10.4 )

#

Code Analysis

Introduces MAX_TOKEN_LEN = 4096 to LegalitySyllableTokenizer and raises ValueError on oversized tokens.

Bounds current_onset growth and reversal to max_onset length, reducing time complexity from O(n^2) to O(n * max_onset).

#

Exploit Details

GitHub Security Advisory : Public advisory containing proof-of-concept verification script demonstrating quadratic CPU growth on all-vowel strings.

#

Mitigation Strategies

• Upgrade NLTK to version 3.10.4 or higher to incorporate algorithmic linearization and token length limits.
• Enforce strict maximum token length validation at the application boundary prior to invoking NLP tokenization methods.
• Implement exception handling for ValueError around tokenization calls to gracefully handle rejected oversized tokens.
• Configure reverse proxies and API gateways to restrict maximum HTTP request payload sizes.

Remediation Steps:

1. Update project dependency specifications (e.g., requirements.txt, pyproject.toml, or Pipfile) to require nltk>=3.10.4.
2. Run 'pip install --upgrade "nltk>=3.10.4"' across all affected virtual environments and container images.
3. Audit application code calling LegalitySyllableTokenizer.tokenize() to ensure ValueError exceptions are caught and handled appropriately.
4. If immediate upgrading is blocked, implement a wrapper function that rejects or truncates tokens exceeding 1,000 characters before passing them to LegalitySyllableTokenizer.

#

References

Read the full report for GHSA-GPWC-27CW-RH9R on our website for more details including interactive diagrams and full exploit analysis.

source & further reading

dev.to — original article How to add an MCP server to Cursor (mcp.json, examples & fixes) Turkish Legal Algorithms in TypeScript: Modeling Law No. 7596 (MKK FIFO & PBLK), Balthazard Disability, and Law No. 6772 Anthropic retires claude-sonnet-4-5 on 30 November 2026: what to change