ios-hig-design-guide

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

• Third-party content exposure detected (high risk: 0.75). Yes—at runtime the required workflow runs scripts/sync_apple_hig_sources.py, which fetches Apple HIG JSON from public endpoints and writes readable text dumps (references/apple-hig-ios-fulltext.md / references/apple-hig-ios-curated.md) that the agent then uses as LLM context; this is public web content ingestion (outsider-authored relative to the operating user).

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

• Potentially malicious external URL detected (high risk: 0.90). The sync script fetches Apple HIG JSON at runtime (e.g. https://developer.apple.com/tutorials/data/index/design--human-interface-guidelines and https://developer.apple.com/tutorials/data{path}.json) and builds the curated/reference files that the skill uses as its source of truth, so external content directly controls the agent's outputs.