linkedin-account-research

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

• Third-party content exposure detected (high risk: 0.85). The skill’s runtime workflow calls UnifAPI endpoints that fetch public LinkedIn company pages/posts/jobs and visible employee profiles (e.g., linkedin/companies/{slug}/posts, linkedin/companies/{slug}/jobs, linkedin/companies/{slug}/people, linkedin/users/{username}), which are outsider-authored free text (company/employee public content) ingested into the agent’s LLM context for summarization.