evaluate-existing-solutions

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

• Third-party content exposure detected (medium risk: 0.30). The skill’s runtime workflow explicitly instructs browsing and using “live external candidates” evidence (e.g., external documentation/source/changelogs and related pages), meaning outsider-authored free text can be ingested during external research without any guaranteed need to preselect a specific item.