x-post

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

• Third-party content exposure detected (high risk: 0.85). In x-check.md Phase 2, the skill navigates to outsider-authored X post/reply pages (other users’ content) via Playwright and extracts reply texts from the accessibility tree, which are then included in the agent’s LLM context (e.g., “Save 3-5 interesting reply texts as samples”).