devops

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

• Third-party content exposure detected (high risk: 0.90). The skill's browser-rendering examples (references/browser-rendering.md) show Puppeteer navigating to external sites (e.g., page.goto('https://news.ycombinator.com') and page.goto(message.body.url)) and feeding the scraped HTML into an AI extractor/queue workflow, so untrusted public web content is ingested and can directly influence agent actions.

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

• Potentially malicious external URL detected (high risk: 0.90). The browser-rendering examples fetch external pages at runtime (e.g., https://news.ycombinator.com) and then inject the fetched page.content() into Ai.run as model input, so remote content fetched during execution can directly control prompts.