Fetching the paper…

ShadowCode: Towards (Automatic) External Prompt Injection Attack against Code LLMs · Around