kube-medic

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

• Third-party content exposure detected (high risk: 0.80). The skill explicitly fetches and parses user-generated cluster content (pod logs, events) via kubectl (see scripts/kube-medic.sh cmd_pod and cmd_sweep and the SKILL.md / SECURITY.md notes stating "Pod logs ... are passed to the LLM for analysis"), so untrusted runtime log/event text can be read and used to drive diagnoses and recommended write actions.