CVE-2026-115802: InternScience InternAgent LoRA Loading experiment.py train inclusion of functionality from untrusted control sphere
A weakness has been identified in InternScience InternAgent 1.5.0. This affects the function train of the file tasks/AutoChem/code/experiment.py of the component LoRA Loading. This manipulation of the argument basemodelnameorpath causes inclusion of functionality from untrusted control sphere. The attack may be initiated remotely. The exploit has been made available to the public and could be used for attacks.
Affected Software
Event History
Frequently Asked Questions
What attacker access is required to exploit this issue?
The issue can be initiated remotely and requires no privileges, but it does require user interaction. Exploitation involves manipulating the base_model_name_or_path argument supplied to the train function.
Which deployment component should be prioritized for investigation?
Prioritize systems using the LoRA Loading functionality in tasks/AutoChem/code/experiment.py, specifically the train function. The affected version identified is InternScience InternAgent 1.5.0.
Is exploit code available?
Yes. A public exploit has been made available, increasing the likelihood that the issue could be used in attacks.