cloud-infra-detector

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

• Third-party content exposure detected (high risk: 0.80). The skill's SKILL.md explicitly ingests and parses public/untrusted third-party signals — "ip_signals", "dns_signals" (CNAMEs), "http_signals" (headers), "tls_signals" (cert issuers), and "repository_signals" (IaC files) — and uses them in detection logic to drive provider/platform decisions, so these external contents can materially influence agent behavior.