speech-pathology-ai

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

• Third-party content exposure detected (high risk: 0.70). The skill's reference code (e.g., references/ai-models.md and references/mellifluo-platform.md) calls transformers.from_pretrained and Wav2Vec2Processor.from_pretrained (e.g., "facebook/wav2vec2-xls-r-300m" and "vitouphy/…") to fetch public Hugging Face models and then uses those model outputs to drive scoring, feedback, and next-practice decisions, so untrusted third‑party content can materially influence agent behavior.