CVE-2026-10036
Received Received - Intake

Arbitrary Code Execution in SpeechBrain via Unsafe YAML Parsing

Vulnerability report for CVE-2026-10036, including description, CVSS score, EPSS score, affected products, exploitability, helpful resources, and attack-flow context.

Publication date: 2026-08-27

Last updated on: 2026-08-27

Assigner: VulnCheck

Description

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.

CVSS Scores

EPSS Scores

Probability:
Percentile:

Meta Information

Published
2026-08-27
Last Modified
2026-08-27
Generated
2026-08-28
AI Q&A
2026-08-28
EPSS Evaluated
N/A
NVD

Affected Vendors & Products

Showing 1 associated CPE
Vendor Product Version / Range
speechbrain speechbrain to 1.1.1 (exc)

Helpful Resources

Exploitability

CWE
CWE Icon
KEV
KEV Icon
CWE ID Description
CWE-502 The product deserializes untrusted data without sufficiently ensuring that the resulting data will be valid.

Attack-Flow Graph

AI Quick Actions

Instant insights powered by AI
Executive Summary

This vulnerability in SpeechBrain before version 1.1.1 allows arbitrary code execution through a crafted CKPT.yaml checkpoint metadata file. The issue occurs because the software uses PyYAML's unsafe loader to parse these files during candidate enumeration. Attackers can embed malicious Python object construction tags like !!python/object/apply in the CKPT.yaml file. When the Checkpointer.recover_if_possible() function processes this file, it executes the embedded code, even if the malicious checkpoint is never selected for recovery.

Detection Guidance

To detect this vulnerability, inspect any CKPT.yaml files in your SpeechBrain checkpoint directories for suspicious YAML tags like !!python/object/apply. Use commands such as 'find / -name CKPT.yaml -type f -exec grep -l "!!python/object" {} \;' to search for malicious patterns. Also check for unsafe YAML loader usage in code with 'grep -r "yaml.Loader" /path/to/speechbrain/'.

Impact Analysis

If you use SpeechBrain versions before 1.1.1, an attacker could exploit this flaw to execute arbitrary code on your system. This could lead to unauthorized access, data theft, or system compromise. The vulnerability is triggered simply by placing a malicious CKPT.yaml file in the checkpoint path, making it a significant risk for users who rely on SpeechBrain for speech processing tasks.

Compliance Impact

This vulnerability could lead to unauthorized code execution, potentially resulting in data breaches or unauthorized access to sensitive information. For organizations subject to GDPR or HIPAA, such incidents may violate compliance requirements, leading to legal penalties, reputational damage, and loss of customer trust. The high CVSS score indicates the severity of the risk.

Mitigation Strategies

Immediately upgrade SpeechBrain to version 1.1.1 or later. Replace any unsafe YAML loader usage with yaml.SafeLoader in your codebase. Remove any CKPT.yaml files containing suspicious Python object construction tags like !!python/object/apply.

Chat Assistant

Ask questions about this CVE
Hi! I’m here to help you understand CVE-2026-10036. Ask me anything about the vulnerability, its impact, or mitigation strategies.
0/70

EPSS Chart