keywords

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

• Third-party content exposure detected (high risk: 1.00). SKILL.md Research process Step 3 explicitly requires the agent to "validate with web search" — including checking existing TikTok content, estimated search interest, and competition level — so the agent will fetch and interpret public social media/web content (untrusted, user-generated) that can materially change keyword decisions.