research-four-steps

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

• Third-party content exposure detected (high risk: 1.00). Flagged because the skill's required workflow (SKILL.md, E — 可执行步骤, step 1) explicitly instructs the agent to "在Google搜索产品关键词和竞品" and to read all existing materials, which means it fetches and interprets open/public web content (including user-generated/untrusted pages) that can directly influence subsequent research and writing actions.