cmo-marketing

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly commissions sub-skills to monitor and act on public, user-generated sources (e.g., the System Overview lists "GitHub, Reddit, Discord, LinkedIn Groups" and Phase 2/Phase 5 require commissioning community-growth with a "Reddit component" and customer-intel for competitor signals), which indicates the agent will ingest and act on untrusted third‑party content that could contain indirect prompt injections.