Arbitrary File Write in Spring AI via Malicious ONNX Model

Executive Summary

This vulnerability in Spring AI allows a local attacker on a multi-user system to pre-create a predictable cache directory path and plant a malicious ONNX model file. This could lead to unauthorized model execution or data compromise.

Useful 0 Not Useful 0

Detection Guidance

Check for unexpected ONNX model files in Spring AI cache directories. Inspect paths like ~/.cache/spring-ai or /tmp/spring-ai for suspicious files. Verify file permissions and ownership of cache directories to detect unauthorized pre-creation.

Useful 0 Not Useful 0

Impact Analysis

An attacker could exploit this to replace a legitimate ONNX model with a malicious one, potentially executing unauthorized code or accessing sensitive data. The impact is high on integrity but does not affect confidentiality or availability.

Useful 0 Not Useful 0

Compliance Impact

This vulnerability could potentially impact compliance with GDPR and HIPAA by allowing unauthorized model execution or data compromise on multi-user systems. The integrity impact (I:H) suggests that malicious model files could alter data or system behavior without detection, which may violate integrity requirements in these regulations.

Useful 0 Not Useful 0

Mitigation Strategies

Upgrade to fixed versions of Spring AI: 2.0.1 (OSS), 2.0.0.1 (Enterprise-only), 1.1.9 (Enterprise-only), or 1.0.10 (Enterprise-only). No additional mitigation steps are required beyond upgrading.

Useful 0 Not Useful 0