A Claude Code plugin for managing and troubleshooting GitHub Actions Runner Controller (ARC) runners on Kubernetes clusters.
ARC Runner Manager provides comprehensive tools for deploying, monitoring, and maintaining GitHub Actions Runner Controller (ARC) infrastructure on any Kubernetes cluster (microk8s, GKE, EKS, AKS, and more).
- Deploy runners — Initialize new ARC runners with GitHub App credentials and Helm
- Diagnose issues — Troubleshoot connection failures, version mismatches, and pod problems
- Scale runners — Adjust min/max runner counts and resource configurations
- Manage credentials — Rotate GitHub App credentials and update authentication
- Production validation — Pre-deployment safety checks and best practices for production environments
- Version management — Handle upgrades, rollbacks, and version alignment between controller and runners
Add the marketplace:
/plugin marketplace add full-stack-biz/claude-arc-runner-manager
Install the plugin:
/plugin install arc-runner-manager@full-stack-biz
Invoke the skill in Claude Code:
/arc-runner-manager
The skill provides three primary workflows:
Diagnose unhealthy runners or connection failures. Includes step-by-step debugging for:
- Runner status and pod health checks
- Controller and listener log analysis
- Root cause identification (auth errors, version mismatches, resource constraints)
- Remediation steps from the troubleshooting guide
Deploy new ARC runners or reinstall existing ones. Covers:
- Prerequisite verification (Kubernetes, Helm, GitHub App setup)
- Secret creation with GitHub App credentials
- Helm-based runner deployment
- Verification in both Kubernetes and GitHub
Adjust runner settings, scaling, and Helm values. Includes:
- Viewing current configurations
- Modifying runner scaling (minRunners/maxRunners)
- Updating labels, resource constraints, and other settings
- Applying changes via Helm upgrade
The plugin includes comprehensive reference guides:
- Deployment Guide — Complete deployment procedures, credential management, version upgrades, and safe upgrade procedures for production
- Troubleshooting Guide — Common errors, diagnosis steps, and remediation for connection failures, version mismatches, pod crashes, and listener issues
- Configuration Reference — All autoscalingrunnersets fields and Helm chart options
- Production Validation — Pre-deployment safety checks, validation scripts, and production-ready checklist
The plugin is organized as follows:
.claude-plugin/
├── plugin.json # Plugin manifest
├── marketplace.json # Marketplace registration
└── skills/arc-runner-manager/
├── SKILL.md # Main skill with workflows
└── references/ # Detailed guides
├── deployment-guide.md
├── troubleshooting-guide.md
├── configuration-reference.md
└── production-validation.md
- Version alignment — Controller (arc-systems) and runner (arc-runners) chart versions must match to avoid silent failures
- Credential security — GitHub App private keys are stored only in Kubernetes secrets; never logged or exposed
- Production validation — Always use the production validation checklist before deploying to production
- Backup procedures — Always backup Helm values before major operations:
helm get values <release> -n arc-runners > backup.yaml
The skill uses:
- Bash — kubectl, helm, and microk8s commands (read-only by default; write operations require explicit user approval)
- Read — View manifests, configurations, and logs
- Edit/Write — Modify configurations (requires confirmation for production changes)
- Kubernetes cluster with kubectl access
- Helm 3+ installed
- GitHub App with proper permissions:
administration:readactions:readrunners:write
- ARC Controller running in the cluster (arc-systems namespace)
Full-stack.biz OU — https://github.com/full-stack-biz
MIT
For issues, documentation, or contributions, visit the repository.