browserless

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

• Third-party content exposure detected (high risk: 0.90). The SKILL.md explicitly instructs the agent to fetch and render arbitrary public URLs (e.g., via the /scrape, /content, /function endpoints with examples like news.ycombinator.com and duckduckgo.com), so untrusted third‑party web content is ingested and can directly influence scraping/execution and subsequent actions.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill makes runtime POST calls to the browserless service (e.g., https://production-sfo.browserless.io/function?token=...) which executes user-supplied JavaScript on the remote service (remote code execution) and the skill depends on that endpoint for core functionality.