PlugWiki

azure-aigateway Skill for Claude: Set Up Azure API Management as an AI Gateway

Makes Claude set up Azure API Management as an AI gateway with token limits, caching, content safety and tool rate limits.

Installs
572.1K
on skills.sh
Stars
1.6K
on GitHub
Files
8
besides SKILL.md
Updated
Oct 2026
last repo commit

azure-aigateway helps Claude configure Azure API Management (APIM) as a gateway in front of AI models, MCP tools and agents. It covers token limits, semantic caching, content safety, load balancing, token metrics and rate limits for MCP tools. It can also add an Azure OpenAI or Foundry backend and test the gateway.

It works with az commands and policy XML. To create the APIM service itself, the SKILL.md sends you to azure-prepare. Several steps create backends, assign roles and read subscription keys, so check each command.

Good for

  • Platform teams governing access to Azure OpenAI and Foundry models
  • Teams adding content safety and jailbreak detection in front of models
  • Owners cutting LLM cost with token limits and semantic caching

Install azure-aigateway

Claude Code

  1. 1

    Copy this prompt.

    Install the azure-aigateway skill in Claude Code. Run this command:
    
    /plugin install azure@claude-plugins-official
    
    Then check that it loaded and tell me how to use it.
  2. 2

    Paste it into Claude Code. It shows the change and asks before running it.

Requires Azure CLI signed in.

Pick one route, not both. The plugin installs all the Azure skills and wires in the Azure and Foundry tool servers they call. The skills CLI route installs only the skills named in the command, and does not wire in the Azure or Foundry tool servers, so enable the Azure MCP server separately if you want the tools the skills call. Sign in with az login first.

From the project README. Source.

What's in the skill

When it loads

Configure Azure API Management as an AI Gateway for AI models, MCP tools, and agents. WHEN: semantic caching, token limit, content safety, load balancing, AI model governance, MCP rate limiting, jailbreak detection, add Azure OpenAI backend, add AI Foundry model, test AI gateway, LLM policies, configure AI backend, token metrics, AI cost control, convert API to MCP, import OpenAPI to gateway.

Sections in SKILL.md

  1. When to Use This Skill
  2. Quick Reference
  3. Get Gateway Details
  4. Test AI Endpoint
  5. Common Tasks
  6. Add AI Backend
  7. Apply AI Governance Policy
  8. Troubleshooting
  9. References
  10. SDK Quick References

Key facts

Author
Microsoft
Last commit
Oct 30, 2026

What it includes

  • Token limits, semantic caching, load balancing, token metrics

    Policies include azure-openai-token-limit, semantic-cache-lookup and store, and emit-token-metric.

    Runs code · Model governance

  • Rate-limits MCP tools and converts an API to MCP

    Uses rate-limit-by-key to protect tools.

    Runs code · Tool governance

  • Content safety and jailbreak detection

    Uses the llm-content-safety policy.

    Write · Agent governance

  • Reads the gateway URL, backends and subscription key

    Runs az apim show, backend list and subscription keys list.

    Runs code · Get gateway details

  • Calls a model through the gateway with curl

    Sends a test chat completion with your subscription key.

    Runs code · Test AI endpoint

  • Creates a backend and grants the gateway access to the model

    Runs az apim backend create and az role assignment create for the Cognitive Services User role.

    Runs code · Add AI backend

From the official docs

Permissions

High risk

Claude can change live gateway policy, assign roles and read subscription keys. Use a test API Management instance and review each command before you approve it.

What it can do

  • Read your data
  • Create and change things
  • Delete things
  • Read-only mode

Watch out for

  • It can assign roles and read subscription keys

    Steps run az role assignment create and list APIM subscription keys. Review each command and keep keys out of the chat.

  • It changes live gateway policy

    Policy edits affect real traffic. Try them on a test APIM instance first.

  • skills.sh audit: Snyk warn

    Gen Agent Trust Hub and Socket pass; Snyk is listed as warn. Read the audit on skills.sh before installing.

Technical details
Scopes
  • Creates and changes APIM backends and policies

  • Assigns a role to the gateway's managed identity

  • Reads APIM subscription keys and calls the gateway with them

Questions

What does azure-aigateway do?

It helps Claude configure Azure API Management as an AI gateway for models, MCP tools and agents. That covers token limits, semantic caching, content safety and load balancing.

How do I install it?

Use the install box at the top of this page. In Claude Code the plugin command installs the Azure skills together with the Azure tools they call, so pick it unless you want this one skill only. Run az login before you start.

Does it create the APIM service?

No. The SKILL.md says to use azure-prepare to deploy APIM, then use this skill to configure it.

Does it need an API key?

It needs the Azure CLI signed in. For testing, it uses your APIM subscription key, which it can read with az apim subscription keys list.

Is it safe?

It changes gateway policy and roles, so use a test instance first. skills.sh lists Gen Agent Trust Hub and Socket as pass and Snyk as warn.

Does it work with models other than Azure OpenAI?

The SKILL.md names Azure OpenAI and Azure AI Foundry models, and MCP tools and agents. It lists no other providers.

Does it work in the Claude app, Codex or Cursor?

The Azure README lists GitHub Copilot, Claude Code, Cursor, Codex, Gemini CLI and OpenCode as hosts for the plugin. It does not list the Claude desktop or web app.

Information

updated Nov 2, 2026

Folder
skills/azure-aigateway

microsoft/azure-skillsskills.sh

Sources

Docs, issue trackers and discussions this page is built from.

Show all 3
  1. microsoft/azure-skills README · github.com
  2. azure-aigateway SKILL.md · github.com
  3. azure-aigateway on skills.sh · skills.sh

Updated Nov 2, 2026.