concept-engine

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

• Third-party content exposure detected (high risk: 0.80). The skill's Phase 1c workflow explicitly instructs the agent to fetch competitor/brand content (e.g., get_workspace_competitors, get_brand_by_domain(brandUrl), get_inspo_creatives, get_inspo_brand_context) and to pull transcripts from creatives, which means the agent will ingest and interpret third‑party/public brand content (including user-provided URLs) and use it to influence concept generation.