lp-brief

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

• Third-party content exposure detected (high risk: 1.00). The skill explicitly accepts and ingests "Existing page state (URL/screenshot/code)" and post-launch evidence as required inputs (see SKILL.md Route B and the Evidence-Anchor Agent input contract in agents/evidence-anchor-agent.md), which means the orchestrator/agents will read and interpret arbitrary public URLs and analytics content to drive hypotheses, architecture, and tool handoffs — exposing the agent to untrusted third‑party content that could enable indirect prompt injection.