Skip to content
Back to skills

Devops Router

ASecurity

Routes DevOps tasks to the correct specialist skill within ccc-devops.

  • 6 stars
  • 0 votes
  • 0 copies
  • 2 views
  • Added June 1, 2026
ai-agentsrustgodockerawsterraformgitdevopsci/cdsecurityperformance

Security analysis

A100/100

Scanned June 1, 2026

npx -y skills add KevinZai/commander --skill devops-router --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Devops Router?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Devops Router
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/kevinzai-devops-router/badge)](https://www.skillsdirectory.com/skills/kevinzai-devops-router)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: DevOps Router
description: "Routes DevOps tasks to the correct specialist skill within ccc-devops."
version: 1.0.0
category: routing
parent: ccc-devops
---

# DevOps Router

> Don't know which DevOps skill to use? Start here. Describe what you need and get routed to the right specialist.

## How It Works

1. Parse the user's request for DevOps signals
2. Match against the intent taxonomy below
3. Recommend 1-3 skills in execution order
4. If ambiguous, ask one clarifying question before routing

## Quick Decision Tree

```
What are you trying to do?
│
├── DEPLOYING code?
│   ├── Need zero-downtime strategy? → zero-downtime-deploy
│   ├── General deployment orchestration? → senior-devops
│   ├── Writing deploy runbook? → infra-runbook
│   └── Setting up deploy pipeline? → cicd-pipeline-generator
│
├── BUILDING CI/CD?
│   ├── Full pipeline (lint→test→build→deploy)? → cicd-pipeline-generator
│   ├── Securing existing workflows? → github-actions-security
│   ├── Making workflows reusable across repos? → github-actions-reusable-workflows
│   └── General DevOps pipeline? → senior-devops
│
├── CONTAINERIZING?
│   ├── Writing/optimizing Dockerfile? → docker-development
│   ├── Docker Compose setup? → docker-development
│   ├── Securing container images? → container-security
│   └── Container runtime security? → container-security
│
├── AWS infrastructure?
│   ├── Architecture design / service selection? → aws-solution-architect
│   ├── Lambda / serverless? → aws-lambda-best-practices
│   ├── S3 / object storage? → aws-s3-patterns
│   ├── CloudFront / CDN? → aws-cloudfront-optimization
│   └── IAM / permissions / roles? → aws-iam-security
│
├── MONITORING / OBSERVABILITY?
│   ├── Prometheus setup / metrics collection? → prometheus-configuration
│   ├── Grafana dashboards? → grafana-dashboards
│   ├── Alert rules / PromQL? → promql-alerting
│   ├── APM / Core Web Vitals / RUM? → performance-monitoring
│   └── Incident response procedures? → infra-runbook
│
├── INFRASTRUCTURE AS CODE?
│   ├── Terraform modules / state / workspaces? → terraform-patterns
│   ├── General IaC scaffolding? → senior-devops + terraform-patterns
│   └── Importing existing infrastructure? → terraform-patterns
│
├── NETWORKING?
│   ├── VPC / subnets / security groups? → network-engineer
│   ├── DNS / load balancing? → network-engineer
│   ├── Service mesh / zero-trust? → network-engineer
│   └── Cloud networking architecture? → network-engineer + aws-solution-architect
│
└── SECURITY?
    ├── Container security? → container-security
    ├── IAM / cloud permissions? → aws-iam-security
    ├── CI/CD security? → github-actions-security
    ├── Network security? → network-engineer
    └── Full security audit? → container-security + aws-iam-security + github-actions-security
```

## Intent Taxonomy

### Deployment
| Signal Words | Route To | Confidence |
|-------------|----------|------------|
| deploy, release, rollout, ship, push to prod | `senior-devops` + `zero-downtime-deploy` | High |
| blue-green, canary, rolling update, zero-downtime | `zero-downtime-deploy` | High |
| rollback, revert, recover, incident | `infra-runbook` | High |
| feature flag, dark launch, traffic splitting | `zero-downtime-deploy` | Medium |

### CI/CD
| Signal Words | Route To | Confidence |
|-------------|----------|------------|
| pipeline, workflow, github actions, CI/CD | `cicd-pipeline-generator` | High |
| secrets, OIDC, supply chain, action pinning | `github-actions-security` | High |
| reusable, composite action, workflow_call, shared workflow | `github-actions-reusable-workflows` | High |
| gitlab ci, jenkins, circleci | `cicd-pipeline-generator` | Medium |

### Containers
| Signal Words | Route To | Confidence |
|-------------|----------|------------|
| dockerfile, docker-compose, multi-stage, image size | `docker-development` | High |
| trivy, grype, CVE, vulnerability scan, image hardening | `container-security` | High |
| container, docker, podman | `docker-development` | Medium — clarify build vs. security |

### AWS
| Signal Words | Route To | Confidence |
|-------------|----------|------------|
| architecture, well-architected, cost optimization | `aws-solution-architect` | High |
| lambda, serverless, cold start, layers | `aws-lambda-best-practices` | High |
| s3, bucket, presigned, lifecycle | `aws-s3-patterns` | High |
| cloudfront, cdn, edge, cache invalidation | `aws-cloudfront-optimization` | High |
| iam, policy, role, permission boundary, scp | `aws-iam-security` | High |
| aws, cloud | `aws-solution-architect` | Low — clarify service |

### Monitoring
| Signal Words | Route To | Confidence |
|-------------|----------|------------|
| prometheus, scrape, exporter, recording rule | `prometheus-configuration` | High |
| grafana, dashboard, panel, variable | `grafana-dashboards` | High |
| promql, alert rule, alertmanager, silence | `promql-alerting` | High |
| apm, rum, core web vitals, sentry, datadog | `performance-monitoring` | High |
| monitoring, observability, metrics | `prometheus-configuration` + `performance-monitoring` | Medium — clarify infra vs. app |

### Infrastructure
| Signal Words | Route To | Confidence |
|-------------|----------|------------|
| terraform, module, state, workspace, plan, apply | `terraform-patterns` | High |
| vpc, subnet, security group, nacl | `network-engineer` | High |
| dns, route53, load balancer, alb, nlb | `network-engineer` | High |
| service mesh, istio, envoy, zero-trust | `network-engineer` | High |
| runbook, playbook, incident, escalation | `infra-runbook` | High |

## Combining Skills

Skills compose naturally. Common combinations:

| Workflow | Skills Used |
|----------|------------|
| New service to production | docker-development → cicd-pipeline-generator → zero-downtime-deploy → performance-monitoring |
| Security hardening | aws-iam-security → container-security → github-actions-security → network-engineer |
| Observability stack | prometheus-configuration → grafana-dashboards → promql-alerting → infra-runbook |
| AWS infrastructure | aws-solution-architect → terraform-patterns → aws-iam-security → network-engineer |
| CI/CD overhaul | cicd-pipeline-generator → github-actions-security → github-actions-reusable-workflows → docker-development |

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.