browser

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

• Third-party content exposure detected (high risk: 1.00). This skill (index.js readUrl) uses Puppeteer to navigate to and extract document.body.innerText from arbitrary public URLs (see SKILL.md and README: "openclaw browser read " / "node index.js read "), so it ingests untrusted third‑party web content that could contain instructions and enable indirect prompt injection.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). This skill explicitly navigates to and renders arbitrary, user-supplied URLs at runtime (e.g., https://example.com) using Puppeteer, executing the page's JavaScript and returning document.body.innerText into the agent context, so remote pages can directly control prompts or inject executable code.