Runner Group Management Skill

Runner Group Management

When to Use This Skill

Runner groups are security boundaries. Organize by trust level. Restrict by default. Enforce with workflow controls.

The Risk

Without runner groups, all self-hosted runners are available to all repositories. One compromised repository means access to production runners with elevated permissions. Malicious workflows can target high-value runners for lateral movement.

Implementation

See the full implementation guide in the source documentation.

Techniques

Runner Group Organization Strategies

Organize runners by security requirements, compliance needs, and operational constraints.

flowchart TB
    A["All Repositories"] --> B{"Trust Level"}

    B --> C["Public Repos"]
    B --> D["Internal Repos"]
    B --> E["Private Repos"]

    C --> C1["public-runners group<br/>No self-hosted access"]

    D --> D1["dev-runners group<br/>Non-production access"]
    D --> D2["test-runners group<br/>Isolated test networks"]

    E --> E1["prod-runners group<br/>Production deployment"]
    E --> E2["compliance-runners group<br/>Audited workloads"]
    E --> E3["gpu-runners group<br/>High-cost resources"]

    %% Ghostty Hardcore Theme
    style C1 fill:#f92572,color:#1b1d1e
    style D1 fill:#e6db74,color:#1b1d1e
    style D2 fill:#e6db74,color:#1b1d1e
    style E1 fill:#a6e22e,color:#1b1d1e
    style E2 fill:#a6e22e,color:#1b1d1e
    style E3 fill:#fd971e,color:#1b1d1e

Strategy 1: Trust-Based Organization

Organize by repository trust level and workflow sensitivity.

| Group Name | Trust Level | Repository Access | Network Scope | Credentials | Use Case | | ---------- | ----------- | ----------------- | ------------- | ----------- | -------- | | public-runners | Untrusted | Public repos only | Internet-only | None | Never use self-hosted for public repos | | dev-runners | Low | Development repos | Isolated dev network | Development service accounts | Feature development, testing | | staging-runners | Medium | Staging repos | Staging network | Staging credentials | Pre-production validation | | prod-runners | High | Production repos | Production network | Production OIDC | Production deployments | | compliance-runners | Highest | Compliance-approved repos | Audited networks | Minimal credentials | HIPAA, PCI-DSS, SOC2 workloads |

Best Practice: Never allow self-hosted runners for public repositories. External contributors can submit malicious workflows that execute on your infrastructure.

Strategy 2: Workload-Based Organization

Organize by job type and resource requirements.

| Group Name | Job Type | Repository Access | Resource Profile | Cost Model | Use Case | | ---------- | -------- | ----------------- | ---------------- | ---------- | -------- | | ci-runners | CI/CD | All repos | 8 CPU, 16GB RAM | Standard | Build, test, lint | | deploy-runners | Deployment | Release repos only | 4 CPU, 8GB RAM | Standard | Cloud deployments | | gpu-runners | ML/AI | ML repos only | 16 CPU, 64GB RAM, 1 GPU | High-cost | Model training, inference | | build-runners | Compilation | Build repos | 16 CPU, 32GB RAM | Standard | Large codebases, monorepo builds | | integration-runners | Integration tests | Test repos | 8 CPU, 16GB RAM | Standard | Database integration, API tests |

Best Practice: Isolate high-cost runners (GPU, high-memory) to prevent accidental usage from unauthorized repositories. Monitor for cost anomalies.

Strategy 3: Environment-Based Organization

Organize by deployment environment and protection rules.

| Group Name | Environment | Protection Rules | Approval Required | Network Access | Use Case | | ---------- | ----------- | ---------------- | ----------------- | -------------- | -------- | | dev-runners | Development | None | No | Dev VPC | Rapid iteration | | staging-runners | Staging | Branch protection | No | Staging VPC | Pre-prod testing | | prod-runners | Production | Environment protection | Yes | Prod VPC | Production deployments | | dr-runners | Disaster Recovery | Manual trigger only | Yes | DR VPC | Failover scenarios |

Best Practice: Combine runner groups with GitHub environment protection rules. Require manual approval before jobs execute on production runners.

Strategy 4: Compliance-Based Organization

Organize by regulatory requirements and audit needs.

| Group Name | Compliance Scope | Audit Logging | Data Classification | Network Isolation | Use Case | | ---------- | ---------------- | ------------- | ------------------- | ----------------- | -------- | | pci-runners | PCI-DSS | Full audit logs to SIEM | Cardholder data | Segmented PCI network | Payment processing | | hipaa-runners | HIPAA | Encrypted logs, BAA | PHI | HIPAA-compliant VPC | Healthcare data | | fedramp-runners | FedRAMP | CloudWatch + Splunk | CUI | FedRAMP-authorized VPC | Government workloads | | sox-runners | SOX | Immutable logs | Financial data | Audited network | Financial reporting | | standard-runners | None | Standard GitHub logs | Public/internal | Standard network | Non-regulated workloads |

Best Practice: Dedicated compliance runners with enhanced logging, immutable audit trails, and network segmentation per regulatory requirements.

See reference.md for additional techniques and detailed examples.

Examples

See examples.md for code examples.

Full Reference

See reference.md for complete documentation.

Agent Skills: Runner Group Management

Install this agent skill to your local

Skill Files