brand-ambassador

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

• Third-party content exposure detected (high risk: 1.00). Yes — SKILL.md explicitly instructs the agent "If the user just names a brand, use web search to research it" and to review the brand's social media presence and recent public content, which requires ingesting untrusted public/UGC web content that will be interpreted and directly influence the agent's outputs.