react-native-best-practices

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly documents loading and processing arbitrary remote resources (e.g., ResourceFetcher.fetch and modelSource 'https://...' in references/on-device-ai/references/core-utilities.md and references/on-device-ai/references/reference-models.md) and includes examples that download and forward public URLs for images/audio (e.g., useImageEmbeddings, useObjectDetection, useVAD, useOCR examples that call model.forward or FileSystem.downloadAsync with https URLs and HuggingFace collections), so untrusted third-party content is ingested and can influence model outputs and subsequent tool actions.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.80). The skill includes runtime model downloads (used by ResourceFetcher and useExecutorchModule) from the Hugging Face collections (e.g. https://huggingface.co/software-mansion/collections), which are required external model files fetched at runtime that directly determine the agent's outputs when loaded.