Automation, AI skills, and deployment tools for OpenShift edge engineering. This repo is the OpenShift Edge team's shared toolkit — it provides cluster deployment automation for edge topologies (two-node HA, SNO, LVM), CI and release workflow tooling, and a Claude Code plugin marketplace that gives engineers AI-assisted skills for day-to-day work.
| Use case | Component |
|---|---|
| Two-node HA cluster (arbiter or fencing) | two-node-toolbox/ |
| EC2 dev host or hypervisor | ec2-deploy/ |
| Single Node OpenShift with DU config | sno-deploy/ |
| LVM Operator development workspace | environments/lvm-operator/ |
| Nightly payload health monitoring | payload-monitor/ |
| Multi-repo workspace management | workspace plugin via the marketplace |
| Claude Code plugins for edge workflows | plugins/ via the marketplace |
git clone git@github.com:openshift-eng/edge-tooling.git
cd edge-tooling
make setup-githooksmake setup-githooks installs a pre-commit hook that runs markdownlint on staged .md files.
Prerequisites: Node.js (for markdownlint), Python 3, Bash.
Each deployment tool has its own prerequisites — see the component README for what you need.
From within Claude Code:
/plugin marketplace add openshift-eng/edge-tooling
Then enable whichever plugins you need. See the plugin README for details.
| Plugin | What it does |
|---|---|
| challenge | Adversarial hypothesis reviewer for root cause analysis |
| edge-ic | IC workflow automation — TODOs, status reports, Jira updates |
| edge-ocp-ci | Monitor nightly payload health across edge topologies with AI-enriched analysis |
| edge-ocp-rc | RC testing for OCP edge topologies (TNF, TNA, SNO) — launch Prow jobs, track results, classify failures |
| edge-scrum | Scrum process management — refinement, sprint planning, standups |
| github | GitHub workflow skills — PR automation, labeling, queue management |
| lvms | LVMS release, QE, and operational workflows |
| lvms-ci | LVMS CI automation — failure triage, Prow job analysis |
| mcp-atlassian | Atlassian Jira MCP server for Claude Code |
| microshift-ci | MicroShift CI failure analysis and JIRA bug creation |
| microshift-dev | MicroShift dev tools collection |
| microshift-release | MicroShift release testing automation |
| pr-review | PR lifecycle toolkit — vet findings, triage CodeRabbit reviews, yolo-agent |
| skills-review | Lint SKILL.md files for quality and correctness |
| threat-model | Security threat analysis for OpenShift PRs |
| two-node | Two-node topology workflow automation — RHEL verification, Jira integration |
| workspace | Multi-repo workspace manager for cross-project AI-assisted development |
| Directory | What it does |
|---|---|
| two-node-toolbox/ | Deploy two-node OpenShift clusters (arbiter/fencing topologies) |
| ec2-deploy/ | Spin up EC2 instances for development and hypervisor use |
| sno-deploy/ | Deploy Single Node OpenShift with DU configuration |
| payload-monitor/ | Nightly payload health monitoring for edge topologies |
| environments/lvm-operator/ | Development workspace for the LVM Storage operator |
| Resource | What it covers |
|---|---|
| CLAUDE.md | Repo overview, tool routing, component-level guides |
| CONTRIBUTING.md | Contribution model, code standards, architectural patterns |
| docs/claude/workflows.md | Deployment walkthroughs (EC2 instance to two-node cluster) |
| docs/claude/prerequisites.md | Required credentials and tools |
| Component READMEs | Per-tool setup, prerequisites, and usage |
PRs use the fork model — push to your fork, open a PR against main. See CONTRIBUTING.md for details.