azure-container-registry

What it does

This skill provides expert guidance for Azure Container Registry development, including troubleshooting, best practices, architecture, security, configuration, integrations, and deployment. It is specifically intended for tasks involving ACR features like Tasks, geo-replication, connected registries, Defender scanning, and Kubernetes image pulls.

How it works

• It directs the agent to use a category index to locate relevant sections. For sections with line ranges, it instructs to read the file using read_file with the specified lines. For sections with file links, it uses read_file on the linked reference file.
• It specifies that the agent must fetch documentation content via network access, preferably using mcp_microsoftdocs:microsoft_docs_fetch with the query from=learn-agent-skill, or fall back to fetch_webpage with accept=text/markdown.
• It includes a structured set of categories (Troubleshooting, Best Practices, Decision Making, Architecture & Design Patterns, Limits & Quotas, Security, Configuration, Integrations & Coding Patterns, Deployment) each with topics and associated URLs to Microsoft Learn content.

When to use it

Use when working on Azure Container Registry related development tasks, including ACR Tasks, geo-replication, connected registries, Defender scanning, or Kubernetes image pulls. The skill notes not to use it for unrelated Azure services (e.g., Azure Container Apps, Azure Container Instances, AKS is covered under a separate skill).

What it can touch

• Tools: claude-code, copilot, codex
• Network access is required to fetch documentation via mcp_microsoftdocs:microsoft_docs_fetch or fetch_webpage.
• Reads documentation content from referenced files or URLs as described in the Category Index.

Caveats

• Requires network access to fetch documentation content. If the mcp_microsoftdocs tool is unavailable, a fallback fetch_webpage method is used.
• The skill is categorized under Security and references official Azure container registry topics and best practices. No outcomes are guaranteed beyond guidance and reference material retrieval.