user-research-synthesis

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

• Third-party content exposure detected (high risk: 0.80). The skill's "Input Formats Supported" explicitly instructs users to paste raw interview notes, survey results, support tickets, and app store reviews—untrusted, user-generated or public third‑party content—which the agent is expected to read and interpret to generate recommendations, so that content can materially influence its actions.