blog-cluster

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

• Third-party content exposure detected (high risk: 0.90). The skill's Plan Phase and semantic-clustering steps explicitly call WebSearch and SERP overlap analysis (e.g., "Plan Phase Step 1" and references/semantic-clustering.md) to fetch and read top-10 organic result URLs and related-search content from the open web, which the agent uses to form cluster membership and execution decisions—exposing it to untrusted third‑party webpages that can influence behavior.