What problem does it solve? Security researchers and bug bounty hunters need a structured methodology to assess Kubernetes and Docker targets, where misconfigurations like anonymous API access, exposed kubelets, and unauthenticated etcd can lead to full cluster compromise but are easy to misidentify or over-report. ## Core Features & Use Cases - Phased attack-surface enumeration: Fingerprint K8s ports (6443/10250/10255/2379/8443), check API anonymous access with SelfSubjectRulesReview, and probe kubelet /run and /exec primitives correctly (SPDY/WebSocket streams, not plain POST). - High-impact exploit paths: Covers etcd 2379 unauthenticated Secret dumps, docker.sock host escape, runc container escapes (CVE-2024-21626), API-server-mediated nodes/proxy RCE, SA token audience/expiry abuse, and Ingress-NGINX CVE-2025-1974. - False-positive prevention: Distinguishes 10255 read-only info disclosure from 10250 RCE, explains why anonymous 200 responses are not cluster-admin, and requires decoded Secret bytes or command output as impact proof. - Use Case: During a bug bounty engagement you discover port 10250 open on a target. Use this Skill to enumerate pods, execute a read-only id command via /run, steal the pod SA token, and validate its real RBAC grants before writing the report. ## Quick Start Use the hunt-k8s skill to assess whether the target's exposed Kubernetes API on port 6443 allows anonymous access with meaningful privileges.