CVE-2026-71281
Received Received - Intake

Arbitrary Code Execution in Hugging Face PEFT LoRA-GA and CorDA

Publication date: 2026-08-05

Last updated on: 2026-08-05

Assigner: 309f9ea4-e3e9-4c6c-b79d-e8eb01244f2c

Description
Hugging Face peft's LoRA-GA and CorDA initialization modules (src/peft/tuners/lora/corda.py lines ~102 and ~163, and src/peft/tuners/lora/loraga.py line ~101) call torch.load() on config-specified cache/covariance files without weights_only=True, bypassing peft's own safe-loading wrapper used elsewhere in the codebase. Because torch.load() without weights_only=True performs full pickle deserialization, loading a malicious cache or covariance file (e.g. a shared/downloaded LoRA-GA or CorDA cache) results in arbitrary code execution.

CVSS Scores

EPSS Scores

Probability:
Percentile:

Meta Information

Published
2026-08-05
Last Modified
2026-08-05
Generated
2026-08-05
AI Q&A
2026-08-05
EPSS Evaluated
N/A
NVD
CVE-2026-71281
EUVD
EUVD-2026-53366

Affected Vendors & Products

Showing 3 associated CPEs
Vendor Product Version / Range
hugging_face peft *
pytorch torch *
hugging_face torch *

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 involves Hugging Face's PEFT library, specifically the LoRA-GA and CorDA initialization modules. These modules use torch.load() without the weights_only=True parameter, which allows full pickle deserialization. This means malicious cache or covariance files could lead to arbitrary code execution when loaded.

Detection Guidance

Check for suspicious files in LoRA-GA or CorDA cache directories, particularly files loaded via torch.load() without weights_only=True. Inspect logs for unexpected model behavior or arbitrary code execution during model loading.

Impact Analysis

If you use PEFT with LoRA-GA or CorDA, an attacker could craft malicious files that execute arbitrary code when loaded. This could compromise your system, data, or model integrity. The high CVSS score (8.8) indicates significant risk of confidentiality, integrity, and availability impacts.

Compliance Impact

This vulnerability could lead to unauthorized code execution, potentially violating GDPR's data protection requirements or HIPAA's security rules for protected health information. Organizations using PEFT may face compliance violations if exploited.

Mitigation Strategies

Update PEFT to a patched version that enforces weights_only=True in torch.load() calls. Avoid loading untrusted LoRA-GA or CorDA cache files. Review and restrict file permissions for cache directories.

Chat Assistant

Ask questions about this CVE
Hi! I’m here to help you understand CVE-2026-71281. Ask me anything about the vulnerability, its impact, or mitigation strategies.
0/70

EPSS Chart