diagnose-ssr

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

• Third-party content exposure detected (high risk: 1.00). The skill's SKILL.md explicitly navigates to and loads an arbitrary target URL (see "Capture the target URL" and the agent-browser open "$URL" / agent-browser get text body steps) and then reads and judges the page content to decide PASS/FAIL and suggested fixes, so untrusted third-party page content can directly influence the agent's decisions.