What does this skill do, and when should you use it?
Azure Deploy is for applications whose Azure project and infrastructure are already prepared. It requires `.azure/deployment-plan.md` from azure-prepare and a `Validated` plan status produced by azure-validate. The skill runs AZD, Terraform, or Azure deployment commands through a checklist, recovery, and verification workflow. It fits development and operations teams shipping existing applications to Azure.
Reads .azure/deployment-plan.md and verifies both the Validated status and populated validation proof; runs the mandatory pre-deploy checklist and selected deployment recipe; executes azd up, azd deploy, terraform apply, or Azure deployment commands; configures SQL managed identity and applies EF migrations when applicable; handles recipe-specific errors, verifies endpoint accessibility, queries Azure for live RBAC role verification, and reports deployed endpoints as fully qualified HTTPS URLs.
- A developer has finished preparing an Azure application and infrastructure and needs to run `azd up`.
- A team wants to push updated code to an existing Azure deployment with `azd deploy`.
- An operations engineer needs to execute Terraform apply or Bicep/Azure deployment after the preparation workflow is complete.
- An application with previously created API Management or AI Gateway infrastructure needs to be deployed.
- A Container Apps project using managed identity with ACR needs an AcrPull RBAC propagation check before deployment.
How do you install this skill?
- This is a static, low-confidence assessment; no Azure command or deployment was executed.
- Provisioning and deployment create external resources and may affect permissions and costs; confirm subscription, region, resource scope, cost, and recovery options before each run.
- The workflow depends on azure-prepare, azure-validate, Azure CLI, AZD, MCP tools, and repository reference files; missing or mismatched versions may interrupt execution.
- No evidence is provided for Chinese-language support or mainland-China network reachability; validate access separately when Azure, Microsoft Learn, or other overseas services are required.
- Shell / CLI
- Network access
- Local filesystem
- MCP Server
Azure Developer CLI (azd)Azure CLI (az)TerraformAzure subscription access
Run npx skills add microsoft/skills and select the required skills in the wizard. Skills are installed in the selected agent directory, such as .github/skills/. The supplied README does not document how to select this exact plugin skill during the wizard.
tmp="$(mktemp -d)"
git clone --depth 1 https://github.com/microsoft/skills.git "$tmp"
mkdir -p ~/.claude/skills
cp -R "$tmp/.github/plugins/azure-skills/skills/azure-deploy" ~/.claude/skills/
rm -rf "$tmp"Generated from the source repository and skill path; it copies only this skill's folder. If the author's install steps above differ, follow those first. To scope it to one project, replace ~/.claude/skills with that project's .claude/skills.
How do you use this skill?
Once installed, send your agent any of these to trigger it:
- run azd up,
- push to production,
- publish to Azure.
First run azure-prepare and ensure .azure/deployment-plan.md exists. Then run azure-validate and confirm the plan status is Validated with populated Validation Proof. Trigger this skill with a request such as “run azd up,” “execute deployment,” “push to production,” or “publish to Azure.” Do not use it to create applications, generate infrastructure, or bypass validation.
What are this skill's strengths and limitations?
- Keeps deployment execution separate from application and infrastructure preparation.
- Covers AZD, Terraform, and Azure deployment commands with error handling and success verification.
- Enforces preparation, validation, pre-deployment, post-deployment, and RBAC checks.
- Requires endpoint reporting with complete HTTPS URLs.
- Cannot create applications, generate infrastructure code, or replace azure-prepare.
- Stops when the deployment plan, Validated status, or validation proof is missing.
- Requires Azure CLI, Azure Developer CLI, Terraform, Azure access, and referenced MCP tooling.
- The supplied material does not show independent tests or the detailed contents of individual deployment recipes.
How does this skill compare with similar options?
Side by side with related skills; every score comes from the same FSRS standard.
| Skill | FS score | Stars | Last updated | License |
|---|---|---|---|---|
| Azure Deployment Runner this page ✓ Microsoft · Official | 54 · Use with care | ★ 2.7k | 3mo ago | MIT |
| Azure Deployment Validator ✓ Microsoft · Official | 46 · Use with care | ★ 2.7k | 3mo ago | MIT |
| Azure Deployment Preparation ✓ Microsoft · Official | 45 · Not recommended | ★ 2.7k | 3mo ago | MIT |
| Azure Enterprise Infrastructure Planner ✓ Microsoft · Official | 48 · Use with care | ★ 2.7k | 3mo ago | MIT |
| Azure Python App Service Deployer ✓ Microsoft · Official | 45 · Not recommended | ★ 2.7k | 3mo ago | MIT |
Unlike azure-prepare, this skill executes deployment only. Unlike azure-validate, it does not perform validation and cannot manually change the plan status to Validated.
How did FollowSkills review this skill?
The skill requires azure-prepare and azure-validate, confirms subscription and location, requires ask_user for destructive actions, and recommends managed identities, scoped RBAC, and no hardcoded credentials. Points are deducted because deployment, provisioning, and potentially costly external cloud effects are broad, rollback is not fully specified, and some flows still use --no-prompt without a complete cost-confirmation and recovery boundary.
The documented steps, recipes, RBAC propagation polling, error tables, and verification flow provide useful operational guidance. Points are deducted because this is static evidence only: key deployment paths, tool availability, reference consistency, and infrastructure combinations were not executed or reproduced, and several failures still shift remediation work to the user.
The audience, prepared-application boundary, triggers, and non-fit creation scenarios are clear, with coverage for AZD, CLI, Bicep, Terraform, CI/CD, and multiple Azure services. Points are deducted because there is no evidence for Chinese-language support, mainland-China network reachability, or broader non-Azure boundaries; high-impact triggers such as “push to production” may still be semantically overbroad.
The skill is layered into metadata, prerequisites, triggers, rules, steps, recipes, and references, and includes MIT licensing, author, and version metadata; repository context provides a license, README, CI, and security-reporting path. Points are deducted because the skill lacks a changelog, explicit maintenance owner or update path, complete installation/tool dependency notes, and relies on implicit context from other skills or external references.
The skill clearly targets deployment of prepared Azure applications and covers execution, verification, endpoint reporting, RBAC checks, and common recovery paths, so the core value is plausible. The static calibration caps this score; points are deducted because there is no execution evidence or representative verified output, and success depends on prerequisite skills, live Azure state, CLI/MCP availability, and project structure.
The documentation supplies concrete commands, conditions, polling logic, and verification procedures, while repository context shows real CI workflows and test infrastructure. Points are deducted because the supplied CI and tests do not demonstrate coverage of azure-deploy key paths, and this review performed no execution, independent reproduction, cross-source corroboration, or deployment-result audit.
Open a dimension to read why it scored that way
Evidence confidence：Low — Mostly static review, author material or a limited demo; useful for discovery, not high-risk decisions.
See the full review method →