Code Project Skill

SkillDev tools

This skill provides patterns, structure, and quality checks for generating full-stack Azure project implementations.

Available today. Use it from your connected AI after setup.

Connect ahel once, and every AI you use reads what you have installed.

Then ask your AI: use the Code Project Skill skill

About this capability

CSA Copilot: the ultimate content factory for Microsoft Cloud Solution Architects

What this skill tells your AI

The instructions your AI receives, as published by olivomarco/vbd-copilot in skills/code-project/SKILL.md and read by ahel’s review.

This skill provides patterns, structure, and quality checks for generating full-stack Azure project implementations.

Mandatory Work Packages

The ai-implementor conductor produces 8 work packages under outputs/ai-projects/<project-slug>/:

1. Infrastructure (infra/)

Bicep modules with parameter files:

  • Logical module separation: networking, security, compute, data, monitoring
  • Parameter files for dev/staging/prod environments
  • Resource naming convention: kebab-case with environment prefix
  • Tags for cost tracking and governance
  • Security: Key Vault for secrets, managed identities, RBAC, private endpoints

2. Application Code (src/)

All application components:

  • Azure SDK with DefaultAzureCredential for authentication
  • Environment-based configuration (no hardcoded values)
  • Clear separation of concerns
  • Package manifest (requirements.txt, package.json, or .csproj)

3. CI/CD Pipelines (.github/workflows/)

GitHub Actions or Azure DevOps pipelines:

  • Infrastructure deployment workflow
  • Application deployment workflow
  • Appropriate triggers (push, PR, manual dispatch)
  • Secrets via ${{ secrets.X }} - never hardcoded

4. Deploy Script (scripts/deploy.sh)

Single idempotent bash script:

set -euo pipefail
Functions: deploy_infra(), deploy_app(), main()
Flags: --help, --infra-only, --app-only
Params: resource group, location, environment (via args or env vars)
Prerequisites check: az CLI, logged-in session

5. Unit Tests (tests/unit/)

  • Standard framework: pytest (Python), Jest (Node), xunit/nunit (.NET)
  • Cover core business logic, not boilerplate
  • Include test config (conftest.py, jest.config, etc.)
  • Runnable with single standard command

6. Smoke Tests (tests/smoke/)

  • Health check for each deployed endpoint
  • Basic API operations (create/read minimum)
  • Use env vars for endpoint URLs
  • Skip gracefully when endpoints unreachable

7. Validation Script (tests/validate.sh)

Single entry-point for verification:

set -euo pipefail
Functions: validate_infra(), run_unit_tests(), run_smoke_tests(), main()
Flag: --live (enables smoke tests)
Prerequisites check: runtimes, az CLI, test frameworks
Summary: PASS/FAIL at the end

8. README (README.md)

Sections: project overview, prerequisites, env setup, infra deployment, app deployment, quick deploy (deploy.sh), validation (validate.sh), local dev, demo guide (with sample I/O), troubleshooting.

Build Order

infra -> app code -> pipelines -> deploy script -> unit tests -> smoke tests -> validate.sh -> README

Review Process

4 specialist reviewers, each scoped to one area:

  • infra-reviewer-subagent: infra/ directory
  • code-reviewer-subagent: src/ and tests/
  • pipeline-reviewer-subagent: .github/workflows/, scripts/deploy.sh, tests/validate.sh
  • docs-reviewer-subagent: README.md

QA Checks

  • infra_qa_checks.py - Bicep syntax, security patterns, naming, tags
  • pipeline_qa_checks.py - Workflow YAML, secret handling, script structure
  • docs_qa_checks.py - README sections, path accuracy, command correctness

Quality Rules

  • No placeholder text (TODO, TBD, FIXME)
  • No emoji characters
  • No em-dashes (use hyphens)
  • Azure-only services and SDKs
  • All environment variables documented

Signals

GitHub stars
67
Forks
44
Last commit
Jul 2026
Advanced
Catalog kind
skill
Gateway key
code-project
Source
github.com/olivomarco/vbd-copilot