Agent-Skills.md

Agent Skills: DevOps Engineering — Quick Reference

DevOps and platform engineering patterns: Kubernetes, Terraform, GitOps, CI/CD, observability, incident response, and cloud-native ops.

UncategorizedID: vasilyu1983/ai-agents-public/ops-devops-platform

Repository

vasilyu1983/ai-agents-public
vasilyu1983License: MIT
4911

Install this agent skill to your local

pnpm dlx add-skill https://github.com/vasilyu1983/AI-Agents-public/tree/HEAD/frameworks/shared-skills/skills/ops-devops-platform

Skill Files

Browse the full folder contents for ops-devops-platform.

Download Skill

Loading file tree…

frameworks/shared-skills/skills/ops-devops-platform/SKILL.md

Skill Metadata

Name
ops-devops-platform
Description
"DevOps and platform engineering patterns: Kubernetes, Terraform, GitOps, CI/CD, observability, incident response, and cloud-native ops."

DevOps Engineering — Quick Reference

This skill equips teams with actionable templates, checklists, and patterns for building self-service platforms, automating infrastructure with GitOps, deploying securely with DevSecOps, scaling with Kubernetes, ensuring reliability through SRE practices, and operating production systems with strong observability.

Modern baseline (2026): IaC (Terraform/OpenTofu/Pulumi), GitOps (Argo CD/Flux), Kubernetes (follow upstream supported releases), OpenTelemetry + Prometheus/Grafana, supply-chain security (SBOM + signing + provenance), policy-as-code (OPA/Gatekeeper or Kyverno), and eBPF-powered networking/security/observability (e.g., Cilium + Tetragon).

Quick Reference

| Task | Tool/Framework | Command | When to Use | |------|----------------|---------|-------------| | Infrastructure as Code | Terraform / OpenTofu | terraform plan && terraform apply | Provision cloud resources declaratively | | GitOps Deployment | Argo CD / Flux | argocd app sync myapp | Continuous reconciliation, declarative deployments | | Container Build | Docker Engine | docker build -t app:v1 . | Package applications with dependencies | | Kubernetes Deployment | kubectl / Helm (Kubernetes) | kubectl apply -f deploy.yaml / helm upgrade app ./chart | Deploy to K8s cluster, manage releases | | CI/CD Pipeline | GitHub Actions | Define workflow in .github/workflows/ci.yml | Automated testing, building, deploying | | Security Scanning | Trivy / Falco / Tetragon | trivy image myapp:latest | Vulnerability scanning, runtime security, eBPF enforcement | | Monitoring & Alerts | Prometheus + Grafana | Configure ServiceMonitor and AlertManager | Observability, SLO tracking, incident alerts | | Load Testing | k6 / Locust | k6 run load-test.js | Performance validation, capacity planning | | Incident Response | PagerDuty / Opsgenie | Configure escalation policies | On-call management, automated escalation | | Platform Engineering | Backstage / Port | Deploy internal developer portal | Self-service infrastructure, golden paths |

Decision Tree: Choosing DevOps Approach

What do you need to accomplish?
    ├─ Infrastructure provisioning?
    │   ├─ Cloud-agnostic → Terraform or OpenTofu (OSS fork)
    │   ├─ Programming-first → Pulumi (TypeScript/Python/Go)
    │   ├─ AWS-specific → CloudFormation or Terraform/OpenTofu
    │   ├─ GCP-specific → Deployment Manager or Terraform/OpenTofu
    │   └─ Azure-specific → ARM/Bicep or Terraform/OpenTofu
    │
    ├─ Application deployment?
    │   ├─ Kubernetes cluster?
    │   │   ├─ Simple deploy → kubectl apply -f manifests/
    │   │   ├─ Complex app → Helm charts
    │   │   └─ GitOps workflow → ArgoCD or FluxCD
    │   └─ Serverless?
    │       ├─ AWS → Lambda + SAM/Serverless Framework
    │       ├─ GCP → Cloud Functions
    │       └─ Azure → Azure Functions
    │
    ├─ CI/CD pipeline setup?
    │   ├─ GitHub-based → GitHub Actions (template-github-actions.md)
    │   ├─ GitLab-based → GitLab CI
    │   ├─ Enterprise → Jenkins or Tekton
    │   └─ Security-first → Add SAST/DAST/SCA scans (template-ci-cd.md)
    │
    ├─ Observability & monitoring?
    │   ├─ Metrics → Prometheus + Grafana
    │   ├─ Distributed tracing → Jaeger or OpenTelemetry
    │   ├─ Logs → Loki or ELK stack
    │   ├─ eBPF-based → Cilium + Hubble (sidecarless)
    │   └─ Unified platform → Datadog or New Relic
    │
    ├─ Incident management?
    │   ├─ On-call rotation → PagerDuty or Opsgenie
    │   ├─ Postmortem → template-postmortem.md
    │   └─ Communication → template-incident-comm.md
    │
    ├─ Platform engineering?
    │   ├─ Self-service → Backstage or Port (internal developer portal)
    │   ├─ Policy enforcement → OPA/Gatekeeper
    │   └─ Golden paths → Template repositories + automation
    │
    └─ Security hardening?
        ├─ Container scanning → Trivy or Grype
        ├─ Runtime security → Falco or Sysdig
        ├─ Secrets management → HashiCorp Vault or cloud-native KMS
        └─ Compliance → CIS Benchmarks, template-security-hardening.md

When to Use This Skill

Claude should invoke this skill when users request:

  • Platform engineering patterns (self-service developer platforms, internal tools)
  • GitOps workflows (ArgoCD, FluxCD, declarative infrastructure management)
  • Infrastructure as Code patterns (Terraform, K8s manifests, policy as code)
  • CI/CD pipelines with DevSecOps (GitHub Actions, security scanning, SAST/DAST/SCA)
  • SRE incident management, escalation, and postmortem templates
  • eBPF-based observability (Cilium, Hubble, kernel-level insights, OpenTelemetry)
  • Kubernetes operational patterns (day-2 operations, resource management, workload placement)
  • Cloud-native monitoring (Prometheus, Grafana, unified observability platforms)
  • Team workflow, communication, handover guides, and runbooks

Resources (Best Practices Guides)

Operational best practices by domain:

Templates (Copy-Paste Ready)

Production templates organized by tech stack:

AWS Cloud

GCP Cloud

Azure Cloud

Kubernetes

Docker

Kafka

Terraform & IaC

CI/CD Pipelines

Monitoring & Observability

Incident Response

Security

Shared Utilities

Centralized patterns from software-clean-code-standard — extract, don't duplicate:

Related Skills

Operations & Infrastructure:

Security & Compliance:

Software Development:

Optional: AI/Automation (Related Skills):

Cost Governance & Capacity Planning

assets/cost-governance/template-cost-governance.md — Production cost control for cloud infrastructure.

Key Sections

  • Cost Governance Framework — Tagging strategy, budget alerts, anomaly detection
  • Cloud Cost Optimization — Right-sizing, reserved capacity, storage tiering
  • Kubernetes Cost Control — Resource requests/limits, quotas, autoscaler config
  • Capacity Planning — Utilization baseline, growth projections, scaling triggers
  • FinOps Practices — Monthly review agenda, optimization workflow

Do / Avoid

Do

  • Tag all resources at creation time
  • Set budget alerts before hitting limits
  • Review right-sizing recommendations monthly
  • Use spot/preemptible for fault-tolerant workloads
  • Set Kubernetes resource requests on all pods
  • Enable cluster autoscaler with scale-down
  • Document capacity planning assumptions
  • Run blameless postmortems after every SEV1/2

Avoid

| Anti-Pattern | Problem | Fix | |--------------|---------|-----| | No cost tags | Can't attribute spend | Enforce tags in CI/CD | | Dev runs 24/7 | ~70% waste | Scheduled shutdown | | Over-provisioned | Paying for idle capacity | Monthly right-sizing review | | No reservations | On-demand premium | 60-70% reserved coverage target | | Alert fatigue | Real issues missed | SLO-based alerting, tuned thresholds | | Snowflake infra | Unreproducible, undocumented | Everything in Terraform/IaC | | Clickops drift | Config outside IaC | Enforce GitOps reconciliation | | No postmortems | Same incidents repeat | Blameless postmortem for SEV1/2 |

Optional: AI/Automation (AIOps)

AI can assist with analysis and triage, but infrastructure/cost/incident decisions require human approval and an audit trail.

See references/aiops-patterns.md for self-healing systems, automated operations, AI-assisted analysis, and bounded claims.

Operational Deep Dives

See references/operational-patterns.md for:

  • Platform engineering blueprints and GitOps reconciliation checklists
  • DevSecOps CI/CD gates, SLO/SLI playbooks, and rollout verification steps
  • Observability patterns (eBPF), incident noise reduction, and reliability drills

External Resources

See data/sources.json for curated sources organized by tech stack:

  • Cloud Platforms: AWS, GCP, Azure documentation and best practices
  • Container Orchestration: Kubernetes, Helm, Kustomize, Docker
  • Infrastructure as Code: Terraform, OpenTofu, Pulumi, CloudFormation, ARM templates
  • CI/CD & GitOps: GitHub Actions, GitLab CI, Jenkins, ArgoCD, FluxCD
  • Streaming: Apache Kafka, Confluent, Strimzi
  • Monitoring: Prometheus, Grafana, Datadog, OpenTelemetry, Jaeger, Cilium/Hubble, Tetragon
  • SRE: Google SRE books, incident response patterns
  • Security: OWASP DevSecOps, CIS Benchmarks, Trivy, Falco
  • Tools: kubectl, k9s, stern, Cosign, Syft, Terragrunt

Use this skill as a hub for safe, modern, and production-grade DevOps patterns. All templates and patterns are operational—no theory or book summaries.

Trend Awareness Protocol

When users ask recommendation questions about DevOps, platform engineering, or cloud infrastructure, validate time-sensitive details (versions, deprecations, licensing, major releases) against primary sources.

Trigger Conditions

  • "What's the best tool for [Kubernetes/IaC/CI-CD/monitoring]?"
  • "What should I use for [container orchestration/GitOps/observability]?"
  • "What's the latest in DevOps/platform engineering?"
  • "Current best practices for [Terraform/ArgoCD/Prometheus]?"
  • "Is [tool/approach] still relevant in 2026?"
  • "[Kubernetes] vs [alternative]?" or "[ArgoCD] vs [FluxCD]?"
  • "Best cloud provider for [use case]?"
  • "What orchestration/monitoring tool should I use?"

Minimum Verification (Preferred Order)

  1. Check the official docs + release notes linked in data/sources.json for the specific tools you recommend.
  2. If internet access is available, confirm recent releases, breaking changes, and deprecations from those release pages.
  3. If internet access is not available, state that versions may have changed and focus on stable selection criteria (operational fit, ecosystem, maturity, team skills, compliance).

What to Report

After searching, provide:

  • Current landscape: What tools/approaches are popular NOW (not 6 months ago)
  • Emerging trends: New tools, patterns, or practices gaining traction
  • Deprecated/declining: Tools/approaches losing relevance or support
  • Recommendation: Based on fresh data, not just static knowledge

Example Topics (verify with fresh search)

  • Kubernetes versions and ecosystem tools (1.33+, Cilium, Gateway API)
  • Infrastructure as Code (Terraform, OpenTofu, Pulumi, CDK)
  • GitOps platforms (ArgoCD, FluxCD, Codefresh)
  • Observability stacks (OpenTelemetry, Grafana stack, Datadog)
  • Platform engineering tools (Backstage, Port, Kratix)
  • CI/CD platforms (GitHub Actions, GitLab CI, Dagger)
  • Cloud-native security (Falco, Trivy, policy engines)

Fact-Checking

  • Use web search/web fetch to verify current external facts, versions, pricing, deadlines, regulations, or platform behavior before final answers.
  • Prefer primary sources; report source links and dates for volatile information.
  • If web access is unavailable, state the limitation and mark guidance as unverified.