whats-new

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

• Third-party content exposure detected (medium risk: 0.65). The skill’s runtime path can ingest outsider-authored free text when RemoteWhatsNewProvider is selected: it fetches JSON from a runtime-supplied endpoint URL (session.data(from:)), decodes description/title strings, and those strings are rendered into WhatsNewView inside the LLM context via the agent’s UI-generation flow.