apple-docs-index

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

• Third-party content exposure detected (high risk: 0.70). The SKILL.md explicitly instructs the agent to fetch and read documentation from public web endpoints (e.g., the sosumi.ai Markdown mirror at https://sosumi.ai/documentation/{framework} and the Apple DocC JSON at developer.apple.com), so the agent will ingest third-party public content (sosumi.ai in particular) that could influence its decisions and tool use.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill explicitly tells the agent to fetch documentation at runtime from https://sosumi.ai/documentation/{framework} (with a JSON fallback at https://developer.apple.com/tutorials/data/index/{framework}), and that remote Markdown/JSON would be retrieved and injected into the agent's context to guide responses, so the external content can directly control prompts.