メニュー
Platform Engineering discipline expertise (SRE, GitOps, DevSecOps, MLOps, Platform Eng). Use for discipline content, tool recommendations, architecture patterns. Triggers on "SRE", "GitOps", "DevSecOps", "MLOps", "platform engineering", "SLO", "ArgoCD", "Falco".
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
| name | platform-expert |
| description | Platform Engineering discipline expertise (SRE, GitOps, DevSecOps, MLOps, Platform Eng). Use for discipline content, tool recommendations, architecture patterns. Triggers on "SRE", "GitOps", "DevSecOps", "MLOps", "platform engineering", "SLO", "ArgoCD", "Falco". |
Consolidated knowledge for all Platform Engineering disciplines and toolkits. Use when writing, reviewing, or verifying platform content.
| Discipline | Modules | Core Concepts |
|---|---|---|
| SRE | 7 | SLOs, error budgets, toil, incident management, capacity planning |
| Platform Engineering | 6 | IDPs, golden paths, self-service, developer experience, SPACE metrics |
| GitOps | 6 | Declarative config, reconciliation, ArgoCD, Flux, progressive delivery |
| DevSecOps | 6 + guide | Shift-left, CI/CD security, supply chain, runtime, security culture |
| MLOps | 6 | Feature stores, model serving, monitoring, Kubeflow, MLflow |
| AIOps | 6 | Anomaly detection, event correlation, observability AI |
| Release Engineering | 5 | Release pipelines, feature flags, canary, blue-green |
| Chaos Engineering | 5 | Failure injection, game days, steady state hypothesis |
| FinOps | 6 | Cloud cost optimization, showback, rightsizing, spot instances |
| Data Engineering | 6 | Streaming, batch, data mesh, Kafka, Spark on K8s |
| Advanced Networking | 6 | DNS at scale, CDN, WAF, BGP, load balancing, zero trust |
| AI/GPU Infrastructure | 6 | GPU scheduling, vGPU, model serving infra, training clusters |
| Engineering Leadership | 6 | Incident command, postmortems, on-call, ADRs, stakeholders |
Service Level Hierarchy: SLI (what you measure) → SLO (target) → SLA (contract). SLA < SLO always.
Error Budget: 1 - SLO target. At 99.9% = 43.2 min/month. Healthy budget → ship fast. Depleted → freeze + fix.
Four Golden Signals: Latency, Traffic, Errors, Saturation.
Toil: Manual, repetitive, automatable work. Target <50% of SRE time. Measure and eliminate.
Four Principles: Declarative, versioned, pulled automatically, continuously reconciled.
Tool Comparison:
| ArgoCD | Flux | |
|---|---|---|
| UI | Rich web UI | CLI-first |
| Multi-cluster | App-of-apps | Kustomization |
| RBAC | Built-in | K8s native |
| Notifications | Built-in | Controllers |
Pipeline Stages: Source (secrets scan) → Build (SAST, SCA, image scan) → Test (DAST) → Deploy (policy, sign, admit)
Supply Chain: SBOM (know ingredients), Sigstore/Cosign (verify origin), SLSA (provenance levels), admission control (enforce at deploy).
Runtime: Falco (detect), Tetragon (enforce at kernel), KubeArmor (least privilege), NetworkPolicies (segment).
IDP Layers: Infrastructure (Crossplane, Terraform) → Platform (K8s, service mesh) → Developer (Backstage, golden paths)
Maturity Levels: L1 (ad-hoc) → L2 (standardized) → L3 (self-service) → L4 (optimized) → L5 (innovation)
SPACE Framework: Satisfaction, Performance, Activity, Communication, Efficiency.
ML Lifecycle: Data → Features → Train → Validate → Serve → Monitor → Retrain
Drift Types: Data drift (input distribution), concept drift (relationship changes), prediction drift (output shifts)
Key Tools: Kubeflow (pipelines), MLflow (tracking), Feast (features), KServe (serving), Evidently (monitoring)
| Category | Key Tools |
|---|---|
| Observability | Prometheus, OpenTelemetry, Grafana, Loki, Jaeger |
| GitOps & Deploy | ArgoCD, Argo Rollouts, Flux, Helm/Kustomize |
| CI/CD | Dagger, Tekton, Argo Workflows |
| Security | Vault+ESO, OPA, Falco, Tetragon, KubeArmor, Supply Chain |
| Networking | Cilium, Service Mesh (Istio) |
| Scaling | Karpenter, KEDA, Velero |
| Platforms | Backstage, Crossplane, cert-manager |
| Developer Exp | k9s, Telepresence, DevPod, Gitpod |
| ML Platforms | Kubeflow, MLflow, Feast, vLLM, Ray Serve |
| Code Quality | SonarQube, Semgrep, CodeQL, Snyk, Trivy |
| Storage | Rook/Ceph, MinIO, Longhorn |
| K8s Distros | k3s, k0s, MicroK8s, Talos, OpenShift |
| IaC | Terraform, OpenTofu, Pulumi, Ansible, CloudFormation, Bicep |
| Section | Modules | Path |
|---|---|---|
| Foundations | 32 | docs/platform/foundations/ |
| Disciplines | 71 | docs/platform/disciplines/ |
| Toolkits | 96 | docs/platform/toolkits/ |
| Total | 199 |
KubeDojo agent session orientation. Use at the start of every fresh session or when picking up a GitHub issue. Triggers on "cold start", "orient", "session start", "issue-driven".
KubeDojo orchestrator role. Drives the module queue, dispatches authors/reviewers, owns PR hygiene, owns session handoffs. Use as primary role for any standalone session on this repo. Triggers on "orchestrate", "drive the queue", "main session", "standalone session".
End-of-session handoff convention for KubeDojo. MARKDOWN handoff file (AI→AI local agent state) + live-index (.agent/STATUS.md) update — never committed to git. Triggers on "session handoff", "end of session", "wrap up session", "write handoff".
Pick the right KubeDojo agent for a task across ALL activities — write / code / review / research / mechanical, not just review. Activity × lane matrix → agent → model → dispatch command. Roster as of 2026-07-07 (cursor, codex, opus/claude-headless, agy[model-selectable, the Google lane — gemini-cli RETIRED], deepseek, grok via grok-CLI/hermes). Use before any dispatch. Triggers on "which agent", "dispatch", "route to", "who should do this".
Cross-family PR review protocol for KubeDojo. Different model family than the author per docs/review-protocol.md. For ANY agent acting as reviewer (codex, composer-2.5, agy, claude). Triggers on "review PR", "R1 review", "R2 review", "cross-family review".
Write KubeDojo curriculum modules. For ANY agent acting as the T0 author (cursor composer-2.5, codex, deepseek-v4-pro, claude headless). Use when creating modules, writing theory, exercises, quizzes. Triggers on "write module", "create module", "new module", "rewrite module".
