follow-unfollow

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

• Third-party content exposure detected (high risk: 0.70). The SKILL.md shows the agent calling external X API endpoints (e.g., GET /x/users/{username} and follower checks on https://xquik.com/api/v1) to resolve handles and list followers, which are untrusted user-generated social-media data that the agent must read and that directly determines follow/unfollow actions.