skill-creator-ms_skill

Overview

This skill is a practical guide for creating and updating skills that extend AI coding agents to work with Azure SDKs and Microsoft Foundry services. It standardizes structure, authentication patterns, testing, and categorization so new skills are consistent, secure, and maintainable. Use it to produce repeatable, battle-tested skill artifacts for Python and other target languages.

How this skill works

The skill inspects required SDK context, enforces DefaultAzureCredential-based authentication patterns, and prescribes a minimal, ordered template for examples and resources. It defines acceptance criteria and test scenarios that validate import patterns, auth, client initialization, and forbidden anti-patterns. It also prescribes categorization, symlink placement, and a lightweight test harness workflow for deterministic validation.

When to use it

• When creating a new skill for an Azure SDK or Microsoft Foundry service.
• When updating an existing skill to reflect breaking SDK changes or new API patterns.
• When onboarding a new contributor to produce consistent, secure skill content.
• When you need to ensure tests and acceptance criteria cover auth and import anti-patterns.
• When packaging reusable code samples, references, and cleanup examples for agent use.

Best practices

• Require the SDK package name and an authoritative documentation URL before starting.
• Keep examples minimal and focused: show auth, client init, one core workflow, and cleanup.
• Always use DefaultAzureCredential and never hardcode secrets; show env var configuration.
• Design test scenarios that assert expected patterns and forbid common mistakes.
• Organize bundled resources (scripts, references, assets) so large docs load only when triggered.

Example use cases

• Create a Python skill for azure-storage-blob that demonstrates upload, list, and delete with DefaultAzureCredential.
• Update an existing foundry skill to match a new SDK version and adjust import paths and method names.
• Author acceptance criteria and mock test scenarios to catch hardcoded credentials or wrong client imports.
• Package streaming patterns and long-running operation examples into a references file for complex services.
• Categorize a new skill under language and product area and create symlinked entries for discovery.