linkedin-dm

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

• Third-party content exposure detected (high risk: 1.00). The skill programmatically navigates LinkedIn (e.g., https://www.linkedin.com/in/me/ and connections' profile pages), takes snapshots and extracts user-generated profile content and recent posts to personalize messages, exposing the agent to untrusted third-party content from public LinkedIn profiles.