idea

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

• Third-party content exposure detected (high risk: 0.90). The skill's required Research workflow (see "Step 3: Continue Session → 3.3b: Research Session") explicitly instructs the agent to use WebSearch to fetch and analyze open web content including forum posts, Reddit threads, and Stack Overflow, meaning it ingests untrusted, user-generated third‑party content that will influence validation and next actions.