container-security

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The skill instructs the agent to read repository files (Dockerfiles, Kubernetes manifests, Helm charts) and to include "evidence" and specific configuration lines in findings/report, which can cause it to output secret values (e.g., Kubernetes Secret data, env vars, API keys) verbatim — creating an exfiltration risk and no redaction guidance is provided.

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

• Third-party content exposure detected (high risk: 0.70). The skill’s runtime workflow ingests the repository’s Dockerfiles/Kubernetes/Helm YAML it “locate[s]” via glob patterns (e.g., **/k8s/**/*.yaml, **/helm/**/*.yaml, **/Dockerfile*), and those files may contain outsider-authored free text (e.g., comments/annotations/strings in manifests) that the agent reads as LLM-readable content for evaluation.