context-builder

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

• Third-party content exposure detected (high risk: 1.00). The skill's Phase 2 Research explicitly requires using WebSearch and WebFetch (via the Task agent) to gather information from the open web (company websites, news, competitive landscape) which is then used to populate the "About" and "Current State" sections and to tailor prompt questions, so untrusted third‑party content will be read and can materially influence subsequent actions.