competitive-war-room

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

• Third-party content exposure detected (high risk: 1.00). Yes—Phase 2's market-researcher prompt explicitly instructs agents to "Use WebSearch extensively" and to cite sources such as G2, Capterra, Reddit and pricing pages, meaning the skill fetches and interprets public, user-generated web content that can directly influence synthesis and actions.