陌讯skills聚合平台

ops-devops-platform

📁 vasilyu1983/ai-agents-public 📅 Jan 23, 2026
28
总安装量
28
周安装量
#7382
全站排名
安装命令
npx skills add https://github.com/vasilyu1983/ai-agents-public --skill ops-devops-platform

Agent 安装分布

claude-code 20
gemini-cli 17
antigravity 16
cursor 16
trae 15

Skill 文档

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:

Each guide includes:

  • Checklists for completeness and safety
  • Common anti-patterns and remediations
  • Step-by-step patterns for safe rollout, rollback, and verification
  • Decision matrices (e.g., deployment, escalation, monitoring strategy)
  • Real-world examples and edge case handling

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

Navigation

Resources

Shared Utilities (Centralized patterns — extract, don’t duplicate)

Templates

Data

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

GOOD: 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 postmortems after every incident

BAD: Avoid

  • Deploying without cost tags
  • Running dev resources 24/7
  • Over-provisioning “just in case”
  • Ignoring reserved capacity opportunities
  • Disabling scale-down to “avoid disruption”
  • Alert fatigue (too many low-priority alerts)
  • Snowflake infrastructure (manual, undocumented)
  • “Clickops” drift (changes outside IaC)

Anti-Patterns

Anti-Pattern Problem Fix
No tagging Can’t attribute costs Enforce tags in CI/CD
Dev runs 24/7 70% waste Scheduled shutdown
Over-provisioned Paying for unused capacity Monthly right-sizing
No reservations Paying on-demand premium 60-70% coverage target
Alert fatigue Real issues missed SLO-based alerting, tuned thresholds
Snowflake infra Undocumented, unreproducible Everything in Terraform/IaC
No postmortems Same incidents repeat Blameless postmortem for every SEV1/2

Optional: AI/Automation (AIOps)

Note: AI can assist with analysis and triage, but infrastructure/cost/incident decisions require human approval and an audit trail (especially anything destructive or irreversible).

AIOps Capabilities (2026)

Self-Healing Systems:

  • AI-powered anomaly detection to predict failures before they happen
  • Automated remediation flows that trigger rollbacks or config changes
  • Intelligent test selection and risk-based change scoring in CI/CD
  • Causal graph analysis for instant root cause identification

Automated Operations:

  • Unused resource detection and notification
  • Right-sizing recommendation generation
  • Alert summarization and correlation (reduce noise by 90%+)
  • Runbook step suggestions and automated execution

AI-Assisted Analysis

  • Cost trend prediction and anomaly detection
  • Incident pattern identification across services
  • Post-mortem theme extraction
  • Capacity planning predictions

Platform Engineering + AI

Platform teams increasingly embed AI capabilities directly into the platform:

  • Multi-agent orchestration for code generation, security validation, deployment
  • Intelligent defaults and guardrails that scale across teams

Bounded Claims

  • AI recommendations need validation before action
  • Automated deletions require approval workflow
  • Cost predictions are estimates, not guarantees
  • Runbook suggestions need SRE verification
  • Self-healing actions should have human-defined policies and audit trails

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)
GitHub 仓库 ↗ ← 返回陌讯 Skills 聚合平台
本页面由 陌讯 Skills 聚合平台 收录整理，数据定期更新。