clawfirm-test

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 shows the browser-agent snapshot/eval/get text workflow against the Clawfirm WKWebView (and even a testserver example pointing at https://example.com), which requires reading and acting on DOM/text from external or user-rendered pages (untrusted third‑party content) to drive clicks, waits, and tool use.