VAITP Dataset

← Back to the dataset

CVE-2026-10036

Arbitrary code execution via crafted CKPT.yaml loaded with PyYAML unsafe loader.

  • CVSS 8.7
  • 502
  • Input Validation and Sanitization
  • Local

SpeechBrain before 1.1.1 contains an arbitrary code execution vulnerability that allows attackers to execute arbitrary code by supplying a crafted CKPT.yaml checkpoint metadata file parsed with PyYAML's unsafe loader during candidate enumeration in Checkpointer.recover_if_possible(). Attackers can embed malicious Python object construction tags such as !!python/object/apply in any CKPT.yaml file within the configured checkpoint path to trigger code execution during candidate discovery, even if the malicious checkpoint is never selected for recovery.

CWE
502
CVSS base score
8.7
Published
2026-08-27
OWASP
A08 Software and Data Integrity Failures
Orthogonal defect classification
Checking
Code defect classification
Incorrect Check
Category
Input Validation and Sanitization
Subcategory
Insecure Parsing or Deserialization
Accessibility scope
Local
Impact
Arbitrary Code Execution
Affected component
SpeechBrain
Fixed by upgrading
Yes

Solution

Upgrade SpeechBrain to version 1.1.1 or newer, which replaces the unsafe PyYAML loader with a safe loader to fix the arbitrary code execution issue.

Vulnerable code sample

import os
import yaml

class Checkpointer:
    def __init__(self, ckpt_dir):
        self.ckpt_dir = ckpt_dir

    def recover_if_possible(self):
        # VULNERABLE: uses unsafe yaml.load allowing arbitrary object construction
        for entry in os.listdir(self.ckpt_dir):
            if entry.endswith('.yaml'):
                path = os.path.join(self.ckpt_dir, entry)
                with open(path, 'r') as f:
                    meta = yaml.load(f, Loader=yaml.FullLoader)  # unsafe
                if meta.get('recoverable'):
                    return meta
        return None

Patched code sample

import os
import yaml

class Checkpointer:
    def __init__(self, ckpt_dir):
        self.ckpt_dir = ckpt_dir

    def recover_if_possible(self):
        # FIX: use safe_load to prevent arbitrary code execution
        for entry in os.listdir(self.ckpt_dir):
            if entry.endswith('.yaml'):
                path = os.path.join(self.ckpt_dir, entry)
                with open(path, 'r') as f:
                    meta = yaml.safe_load(f)  # safe
                if meta and meta.get('recoverable'):
                    return meta
        return None

Payload

__VAITP_MODEL_REFUSED__

Cite this entry

@misc{vaitp:cve202610036,
  title        = {{Arbitrary code execution via crafted CKPT.yaml loaded with PyYAML unsafe loader.}},
  author       = {Bogaerts, Fr\'ed\'eric and Ivaki, Naghmeh and Fonseca, Jos\'e},
  year         = {2026},
  note         = {VAITP Python Vulnerability Dataset, entry CVE-2026-10036},
  howpublished = {\url{https://netpack.pt/vaitp/vulnerability/CVE-2026-10036/}}
}
Introducing the "VAITP dataset": a specialized repository of Python vulnerabilities and patches, meticulously compiled for the use of the security research community. As Python's prominence grows, understanding and addressing potential security vulnerabilities become crucial. Crafted by and for the cybersecurity community, this dataset offers a valuable resource for researchers, analysts, and developers to analyze and mitigate the security risks associated with Python. Through the comprehensive exploration of vulnerabilities and corresponding patches, the VAITP dataset fosters a safer and more resilient Python ecosystem, encouraging collaborative advancements in programming security.

The supreme art of war is to subdue the enemy without fighting.

Sun Tzu – “The Art of War”

:: Shaping the future through research and ingenuity ::