Flora is a Zarf-based orchestration repository designed to build, harden, and package Platform STAQ (Secure Tooling & AI Quarters). It enables the deployment of a full DevOps and AI development suite into air-gapped or highly regulated environments.
The repository is organized into modular Zarf packages, allowing for a "foundation-first" deployment model.
| Package | Component | Versions & Tools | Description |
|---|---|---|---|
| 01-staq-core | Foundation | Traefik (Chart 42.1.3 / v3.7.16), Gateway API (v1.6.3), Cert-Manager (v1.21.3), Kube-Prometheus-Stack (92.14.1 / Prom v3.15.1, Grafana 13.2.4), Loki (7.5.0 / v3.6.11), Promtail (6.17.1 / v3.5.1) | Mandatory. Ingress, Cert-Manager, Logging, and Observability foundation. |
| 02a-staq-gitlab | Source Control | GitLab (Chart 9.11.14 / v18.11.13 CE), Runner (v18.11.2), PostgreSQL (16.6.0), Redis (7.2.4), MinIO | Self-hosted GitLab instance with integrated container registry, runners, and storage. |
| 02b-staq-factory | DevSecOps | ArgoCD (Chart 10.11.6 / v3.5.6), SonarQube (2026.5.1003 / 26.9.0-community), Nexus (64.2.0 / 3.97.3), Harbor (1.19.4 / 2.15.4) | Tooling factory including GitOps CD, SAST, artifact repository, and container registry. |
| 03-staq-ide | AI Workspace | Code-Server (Chart 3.0.0 / 4.100.0), Ollama (Chart 1.92.0 / 0.40.11), Open WebUI (Chart 17.0.0 / 0.12.0), TinyLlama 1.1B | Web-based VS Code IDE, CPU-based LLM chatbot, and Open WebUI frontend. |
| 04-staq-custom | User Stack | Custom cherry-picked components | A template for cherry-picking specific components for a bespoke deployment. |
The platform uses GitHub Actions to automate the transition from source code to airgap-ready OCI artifacts:
- Model Hardening: (See
packages/03-staq-ide/Dockerfile.model) The pipeline downloads raw GGUF models, performs a security scan, and wraps them into a hardened data-container image - Package Creation: Zarf bundles Kubernetes manifests, Helm charts, and images into OCI artifacts
- OCI Publishing: Artifacts are pushed directly to GitHub Container Registry (
ghcr.io/matteus8/flora)
- Zarf CLI installed locally
- Access to a Kubernetes cluster (K8s or K3s)
echo "<GITHUB_PAT>" | zarf tools registry login ghcr.io -u <GITHUB_USERNAME> --password-stdinThe Core package must be deployed before any other services:
zarf package deploy oci://ghcr.io/matteus8/flora/staq-core:<VERSION> --confirmIncludes the hardened AI model and web-based VS Code environment:
zarf package deploy oci://ghcr.io/matteus8/flora/staq-ide:<VERSION> --confirmTo move this platform into a disconnected environment:
- Pull the artifacts on a connected machine:
zarf package pull oci://ghcr.io/matteus8/flora/staq-core:<VERSION> zarf package pull oci://ghcr.io/matteus8/flora/staq-ide:<VERSION>
- Transfer the resulting
.tar.zstfiles to the airgapped network - Initialize the airgap cluster:
zarf init
- Deploy locally:
zarf package deploy zarf-package-staq-core-amd64-<VERSION>.tar.zst
Note: For custom configurations, refer to the zarf-config.yaml at the root of the repository to set global deployment variables