ai-ml-development

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 1.00). SKILL.md explicitly shows loading public, user-generated data (e.g., datasets.load_dataset("imdb")) and pulling/finetuning models from Hugging Face (from_pretrained + LoRA), so the skill ingests untrusted third‑party content that can materially change model behavior.