funsloth-train
Warn
Audited by Snyk on Feb 16, 2026
Risk Level: MEDIUM
Full Analysis
MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).
- Third-party content exposure detected (high risk: 0.80). The generated notebook templates include explicit steps to "Load Dataset" and use identifiers like dataset_name "mlabonne/FineTome-100k" and FastLanguageModel.from_pretrained(), which implies fetching models/datasets from public hubs (e.g., Hugging Face) and thus exposes the workflow to untrusted, user-generated third-party content that could contain indirect prompt injections.
Audit Metadata