mapkit

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

• Third-party content exposure detected (high risk: 0.80). This skill's SKILL.md "Fetching More Docs" workflow explicitly directs the agent to fetch documentation pages from the public sosumi.ai mirror (e.g., https://sosumi.ai/documentation/mapkit/mkmapview), so the agent will retrieve and interpret open third‑party web content at runtime which can materially 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 instructs at runtime to fetch missing documentation from the sosumi.ai Markdown mirror (e.g. https://sosumi.ai/documentation/mapkit/mkmapview), which would inject external Markdown into the agent’s context and could directly influence prompts/responses.