solve-challenge

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

• Third-party content exposure detected (high risk: 1.00). The skill explicitly instructs the agent to "Fetch everything — Download files, visit URLs, connect to services (nc), read source code" and includes commands like "curl -s http://target/" plus OSINT steps (social media, CTFtime writeups), meaning it will fetch and interpret untrusted public web and user-generated content.