ワンクリックで
deployment-strategy
Document source control, CI/CD platform, environments, deployment process, and quality gates for the application
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。
メニュー
Document source control, CI/CD platform, environments, deployment process, and quality gates for the application
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
Generate Azure infrastructure as Bicep code from planning data
Generate Azure infrastructure as Terraform code from planning data
Generate a Mermaid architecture diagram derived from application components and infrastructure context data
Configure an individual application component by mapping it to an Azure service with SKU, region, and service-specific settings
Read all planning data, propose relevant ADR topics, and generate Architecture Decision Records following the project template
Gather network topology, landing zones, existing resources, and connectivity requirements for Azure infrastructure planning
| name | deployment-strategy |
| description | Document source control, CI/CD platform, environments, deployment process, and quality gates for the application |
This skill documents the end-to-end deployment strategy for the application. It generates infra/application-architecture/deployment-strategy.md which serves as the reference for CI/CD pipeline design and IaC deployment automation.
The generated file will be displayed in the UI when the user runs npx @zureltd/az-infra-harness and navigates to the Application Architecture section.
Run this skill when:
/deployment-strategyCheck whether infra/application-architecture/deployment-strategy.md already exists.
Also scan the codebase for existing CI/CD configuration:
.github/workflows/ — GitHub Actions workflowsazure-pipelines.yml — Azure Pipelines.gitlab-ci.yml — GitLab CIinfra/context/development-context.md — already-captured development context (reuse this!)If existing deployment strategy file found → follow Update Mode (Step 2a). If no existing file → follow Fresh Mode (Step 2b).
Read the existing file and present the current content to the user.
I found your existing deployment strategy. Here's what's currently documented:
[Summary of each section]
What would you like to update? You can modify any section, or say "looks good" to confirm everything is correct.
Wait for the user's response, then make changes. Ask "Anything else to update, or shall I save?" before proceeding to Step 5.
This skill MUST comply with the shared Interaction & Validation Standard:
.opencode/skills/_shared/interaction-validation-standard.md
Important: First check infra/context/development-context.md — if it exists, many answers may already be captured there. Pre-populate what you can from that file and ask the user to confirm rather than asking from scratch.
Use the TodoWrite tool to create a task list at the start with these high-priority items:
Ask one section at a time, wait for the response, then mark the task complete before moving to the next.
Let's start with your source control setup:
1. **Repository URL**: What is the repository URL or name? (e.g., github.com/contoso/customer-portal)
2. **Branching strategy**: What branching strategy do you use? (e.g., GitFlow with main/develop/feature/hotfix, trunk-based development)
3. **Branch protection**: Are there branch protection rules on main/develop? (e.g., require PR, required reviewers, status checks)
If development-context.md exists: "Based on your development context, you're using [platform] with [branching strategy]. Can you confirm the repository URL and branch protection rules?"
Now let's capture your CI/CD setup:
1. **Platform**: What CI/CD platform do you use? (e.g., GitHub Actions, Azure Pipelines)
2. **Workflow structure**: Is there a single pipeline for the whole application, or separate pipelines per component?
3. **Secrets management**: How are secrets (Azure credentials, API keys) passed to the pipeline? (e.g., GitHub Secrets + Azure Key Vault reference, pipeline library variables)
If development-context.md or CI config found: "I can see you're using [platform]. Is there a separate pipeline per component or one shared pipeline?"
What environments does your application have, and how does code flow through them?
For each environment, please describe:
- **Name** (e.g., dev, staging, production)
- **Trigger**: What causes a deployment to this environment? (e.g., push to develop branch, manual approval, merge to main)
- **Approval gate**: Is there a manual approval required? Who approves?
- **Azure subscription/resource group**: Does each environment use a separate subscription?
Common example: dev (auto on PR merge to develop) → staging (manual approval) → production (manual approval + change ticket)
Validation:
Walk me through the numbered steps of your deployment process from code commit to production:
For example:
1. Developer opens PR → CI runs tests and scans
2. PR is reviewed and merged to develop
3. CI/CD auto-deploys to dev environment
4. QA validates in staging after manual approval
5. Release manager approves production deployment
6. CI/CD deploys to production using Bicep/Terraform what-if then apply
Please describe your actual process.
Validation:
What quality gates must pass before a deployment proceeds?
Common gates to describe:
- **Test coverage**: Minimum test coverage threshold (e.g., 80%)
- **Integration tests**: Do integration tests run in the pipeline?
- **Security scanning**: Any SAST/DAST/dependency scanning (e.g., Snyk, SonarQube)?
- **IaC validation**: Is Bicep lint / Terraform validate / plan run before apply?
- **Performance tests**: Any load/performance gates in staging?
Before generating the file, summarize any missing information:
Open Items:
- [Missing Item 1]
- [Missing Item 2]
Resolve all open items. Only proceed when there are zero open items.
Once all sections are complete and validated, create the markdown file with this exact structure:
# Deployment Strategy
## Source Control
[Repository URL. Branching strategy. Branch protection rules.]
## CI/CD Platform
[Platform name. Workflow structure (per-component or shared). Secrets management approach.]
## Environments
[List of environments with trigger, approval gate, and subscription details.]
## Deployment Process
1. [Step 1]
2. [Step 2]
3. [Step 3]
...
## Quality Gates
[Test coverage requirements. Integration test approach. Security scanning tools. IaC validation steps. Performance testing.]
Formatting rules:
#) for the file title: exactly # Deployment Strategy##) for each sectionBefore saving:
# Deployment StrategyTarget location: infra/application-architecture/deployment-strategy.md
Pre-save checks:
infra/application-architecture/ existsError handling:
✅ Created deployment strategy successfully!
📄 File location: infra/application-architecture/deployment-strategy.md
🌐 To view in the UI:
1. Ensure the Az Infra Harness is running: `npx @zureltd/az-infra-harness`
2. Refresh your browser
3. Navigate to the Application Architecture section
4. The deployment strategy card should now show a blue border with a checkmark
The next step is to run /generate-adrs to generate architecture decision records.
Agent: "Let's document your deployment strategy. I can see from your development context that you're using GitHub Actions with GitFlow. Can you confirm the repository URL?"
User: "github.com/contoso/customer-portal"
Agent: "And is there a separate GitHub Actions workflow per component, or one shared workflow?"
User: "Separate workflows per component."
[Conversation continues through all sections]
Agent: "✅ Created deployment strategy successfully!
📄 File location: infra/application-architecture/deployment-strategy.md"
infra/application-architecture/deployment-strategy.mdinfra/context/development-context.md.opencode/skills/_shared/interaction-validation-standard.mdDATA-STRUCTURE.mdinfra/application-architecture/deployment-strategy.md