scanning-kubernetes-manifests-with-kubesec

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

• Third-party content exposure detected (high risk: 0.80). The skill explicitly posts manifests to the public Kubesec API (https://v2.kubesec.io/scan) in scripts/agent.py::_scan_via_api and allows arbitrary kubesec_url in scripts/process.py (using curl), then parses the JSON response and uses scoring/advise/critical fields to make gating/exit decisions and remediation actions—i.e., it ingests untrusted third‑party content that can materially influence behavior.

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

• Potentially malicious external URL detected (high risk: 0.70). The skill includes an installation/CI step that downloads and installs an executable binary from https://github.com/controlplaneio/kubesec/releases/latest/download/kubesec_linux_amd64.tar.gz (curl ... | tar ...), which fetches remote code used at runtime as a required dependency and therefore can execute remote code.