timesfm-forecasting

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

• Third-party content exposure detected (high risk: 0.70). The skill explicitly downloads pretrained model weights from the public Hugging Face Hub (e.g., "google/timesfm-2.5-200m-pytorch" referenced in SKILL.md and via timesfm.from_pretrained / TimesFmCheckpoint), so it ingests content from an open third‑party source which the agent loads and whose outputs directly affect forecasting decisions.