halo-cli

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

• Third-party content exposure detected (high risk: 0.80). The SKILL.md explicitly references a "halo-cli-search: public site search" and the halo search --url flow, indicating the agent will fetch and read public/user-generated site content (posts, pages, comments) from arbitrary public sites which could influence subsequent CLI actions like moderation or content management.