用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/Azure/vscode-aks-tools --skill kickstart-deploy命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
正在显示 SKILL.md
Prepare a release PR for the Azure/vscode-aks-tools VS Code extension. Bumps version in package.json and package-lock.json, creates a new What's New doc, updates SUMMARY.md / release.md / README.md / releasing.md, and removes stale prior-release files. WHEN: 'make a release PR', 'release vscode-aks-tools', 'bump extension version', 'create publish-x.y.z branch', 'prepare release notes for aks tools', 'what's new file for next release'. DO NOT USE FOR: actually publishing to the marketplace (that runs via the internal 1ES pipeline), changelog backfill (CHANGELOG.md is deprecated since 1.6.14), or unrelated repos.
ACR integration for AKS Automatic. Teaches attaching an ACR, image reference conventions (digest pinning, no :latest), and pull-secret-free authentication via the managed identity.
Writing idiomatic, safe, and reviewable Bicep templates for Azure resources.
基于 SOC 职业分类
| name | kickstart-deploy |
| description | Deploy phase playbook — build, push, apply with Azure CLI and kubectl. |
| disable-model-invocation | true |
Deploy using Azure CLI and kubectl. Execute each step via run_in_terminal, confirm between steps with vscode_askQuestions. Never auto-deploy.
Build and push: az acr build --registry <acr> --image <image>:<tag> -f <dockerfilePath> <buildContext>
Use the build context and Dockerfile path from the structure map — never assume repo root (.). For monorepos, build each service from its own context. Tag with a version (e.g. v1.0.0), never :latest.
Get credentials: az aks get-credentials --resource-group <rg> --name <cluster> --overwrite-existing
kubelogin handles AAD auth automatically (verified in Pre-Deploy Check). Never use --admin.
Apply manifests: kubectl apply -f k8s/
Verify: kubectl get pods -n <namespace> and kubectl get services -n <namespace>
If pods not Ready, run kubectl describe pod <name> and kubectl logs <name> to diagnose.
Health-check the running app: don't declare success on pod readiness alone — actually hit the app and compare against the expected response:
kubectl get httproute / kubectl get services: curl -sS -o /dev/null -w "%{http_code}" http://<url>/ (expect 2xx/3xx).kubectl exec <pod> -n <namespace> -- curl -sS localhost:<port>/<health-path>.
If the response isn't what the app should return (wrong status, error body, or logs show a missing entry point), classify as a cluster failure and diagnose before reporting success.Classify failures:
Provide specific az or kubectl fix commands. Offer retry via vscode_askQuestions.
Only mention GitHub Actions if the user asks about CI/CD.