run-advanced-patent-query-ip

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

• Third-party content exposure detected (medium risk: 0.30). At runtime, the skill takes outsider-authored free text from the user-provided Patsnap query and sends it as query_text/derived literature keywords to external Patsnap/literature endpoints, which then return untrusted text fields (titles/abstracts/metadata) that the agent renders into Markdown/HTML.