CVE-2026-126996: SakanaAI text-to-LoRA Checkpoint Loading model_loading.py get_tokenizer path traversal
A vulnerability was determined in SakanaAI text-to-LoRA 0.0.1. Affected by this issue is the function gettokenizer of the file src/hyperllmmodulator/utils/modelloading.py of the component Checkpoint Loading. Executing a manipulation of the argument basemodelnameorpath can lead to path traversal. The attack may be launched remotely. The exploit has been publicly disclosed and may be utilized. The project was informed of the problem early through an issue report but has not responded yet.
Event History
Frequently Asked Questions
What access does an attacker need to exploit this issue?
The vulnerability can be targeted remotely and requires no privileges, but exploitation requires user interaction. An attacker would need to induce a user or process to supply a manipulated base_model_name_or_path value to the affected checkpoint-loading functionality.
Which deployments are affected?
The affected version identified is SakanaAI text-to-LoRA 0.0.1. The issue is in get_tokenizer within src/hyper_llm_modulator/utils/model_loading.py, specifically when handling the base_model_name_or_path argument.
Is public exploit information available?
Yes. A proof of concept has been flagged, and the exploit has been publicly disclosed, so exploitation may be practical.
Is a vendor fix available?
The available information does not identify a fix. The project was reportedly notified through an issue report but had not responded at the time of publication.