fallback-and-resilience

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

• Third-party content exposure detected (high risk: 0.80). The SKILL.md explicitly instructs the agent to access and act on public third‑party websites and services (e.g., "check the airline website directly", "search the airline's website directly", use Skiplagged/Kiwi.com/SerpAPI/web search) and to use those results to decide fallbacks, so untrusted web content can influence tool use and decisions.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly instructs at runtime to pull and run a Docker image (docker pull ghcr.io/borski/sw-fares), which would fetch and external container and execute remote code, so this URL is a runtime dependency that can execute code.