linkedin-outreach

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

• Third-party content exposure detected (high risk: 0.75). Outsider-authored free text from LinkedIn pages (e.g., post/comment text and profile/headline snippets) is extracted at runtime via Playwright page.evaluate(...) in scripts/linkedin_outreach.py and then stored/returned in the RESULT: JSON, which can be ingested into the agent context.