Is russellading/vagrant-kubeadm-kubernetes safe?
Yes. russellading/vagrant-kubeadm-kubernetes is safe to install. Oathe's behavioral security audit gave the vagrant-kubeadm-kubernetes skill by russellading a trust score of 85/100 with 7 findings, none critical or high.
https://github.com/russellading/vagrant-kubeadm-kubernetes
Is russellading/vagrant-kubeadm-kubernetes safe to install?
This repository (russellading/vagrant-kubeadm-kubernetes) is a legitimate Kubernetes practice environment provisioning toolkit intended for CKA/CKAD exam preparation, not an LLM agent skill. Its SKILL.md is empty, eliminating any prompt injection vector. The primary risk is a supply chain dependency on a third-party non-official GitHub account (techiescamp) for a metrics-server manifest that is downloaded and applied without checksum verification during cluster setup. Canary file integrity is confirmed intact; read-access events on sensitive files are attributable to the Oathe audit framework itself.
What security issues were found in russellading/vagrant-kubeadm-kubernetes?
Category Scores
Findings (7)
MEDIUM Third-party metrics-server manifest downloaded and applied -22 ▶
master.sh fetches https://raw.githubusercontent.com/techiescamp/kubeadm-scripts/main/manifests/metrics-server.yaml from a non-official third-party GitHub account (techiescamp) and applies it directly to the Kubernetes cluster with kubectl. If the techiescamp repo is compromised or deleted and re-created with malicious content, any deployment of this Vagrantfile would run attacker-controlled Kubernetes manifests with cluster-admin-equivalent access.
Evidence: kubectl apply -f https://raw.githubusercontent.com/techiescamp/kubeadm-scripts/main/manifests/metrics-server.yaml
LOW Remote manifest application without checksum verification -10 ▶
master.sh and dashboard.sh fetch YAML manifests from raw.githubusercontent.com and apply them via kubectl apply without any SHA or digest verification. The CALICO_VERSION and DASHBOARD_VERSION variables are sourced from settings.yaml and interpolated into URLs. A tampered settings.yaml or a hijacked GitHub release tag could serve malicious manifests.
Evidence: curl https://raw.githubusercontent.com/projectcalico/calico/v${CALICO_VERSION}/manifests/calico.yaml -O && kubectl apply -f calico.yaml; kubectl apply -f "https://raw.githubusercontent.com/kubernetes/dashboard/v${DASHBOARD_VERSION}/aio/deploy/recommended.yaml"
LOW Canary file read events present in audit log -8 ▶
Filesystem and auditd logs show open/access syscalls on /home/oc-exec/.env, .ssh/id_rsa, .aws/credentials, .npmrc, .docker/config.json, and gcloud credentials. Timestamps place first access at 1794409989 (before git clone at 1794409994) and second batch at 1794410010 (end of monitoring). Both batches are consistent with the Oathe audit framework's own initialization and canary-check routines, not repo code. System confirms all files intact.
Evidence: type=PATH msg=audit(1794409989.141:269): name="/home/oc-exec/.env"; type=PATH msg=audit(1794410010.896:1868): name="/home/oc-exec/.env"
INFO SKILL.md is empty — no agent injection surface 0 ▶
The SKILL.md file contains no content. This repository cannot inject instructions into an LLM agent's system prompt because there is nothing to inject. The repository is a Vagrant/Kubernetes provisioning toolkit, not an LLM skill.
Evidence: SKILL.md content is empty string
INFO Clone network activity limited to GitHub 0 ▶
All outbound connections during the clone phase target 140.82.121.4:443 (api.github.com / objects.githubusercontent.com). No connections to unexpected third-party infrastructure were initiated during the clone operation itself.
Evidence: SOCKADDR saddr_fam=inet laddr=140.82.121.4 lport=443 — from git-remote-http
LOW Not a valid LLM skill — empty SKILL.md with privileged shell scripts -5 ▶
This repository has no SKILL.md content and is a Kubernetes cluster provisioning tool. Its shell scripts require root privileges, modify kernel parameters (sysctl, modules), install system packages, and configure networking. If an agent were instructed to execute these scripts in a production environment, it could significantly alter host system configuration.
Evidence: scripts/common.sh: sudo sysctl --system; sudo modprobe overlay; sudo apt-get install kubelet kubectl kubeadm
INFO All honeypot canary files confirmed intact 0 ▶
The Oathe monitoring system confirms no canary file was modified. Read-only access events are attributable to the audit framework's own initialization and final integrity check routines.
Evidence: Monitoring: Canary File Integrity — ✅ All canary files intact (no exfiltration detected via honeypots)
Should I install russellading/vagrant-kubeadm-kubernetes?
Oathe's verdict for russellading/vagrant-kubeadm-kubernetes is SAFE with a trust score of 85/100. Recommendation: Install with caution.