axolotl

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

• Third-party content exposure detected (high risk: 0.90). The skill's documentation (see references/dataset-formats.md) explicitly describes loading datasets from the Hugging Face hub and arbitrary public/local dataset files (e.g., pretraining_dataset: hf_org/name and data_files like A.jsonl), so the agent may ingest and process untrusted, user-generated third‑party content as part of its workflow.