multi-ai-collab

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

• Third-party content exposure detected (high risk: 0.70). This skill explicitly directs use of the Gemini CLI for "web search" and "CVE search" (e.g., "Gemini has web search capabilities" and "can search latest CVEs and security advisories"), so agents in the workflow may fetch and ingest open/public web content (third‑party, user-generated or external advisories) that the orchestrator/agents will read and interpret, creating a risk of indirect prompt injection.