azure-aigateway

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.

Install with Tessl CLI

npx tessl i github:microsoft/github-copilot-for-azure --skill azure-aigateway

What are skills?

1.72x

Review — 93%

Does it follow best practices?

Evaluation — 95%

↑ 1.72x

Agent success when using this skill

Validation — 11 / 11 Passed

Validation for skill structure

SKILL.md

Review

Evals

Evaluation results

100%

24%

AI Gateway Policy Stack for Financial Services

Full AI governance policy composition

Criteria

Without context

With context

Auth before cache

100%

Cache before token limit

100%

Token limit before content policy

100%

Metrics last in inbound

100%

Correct auth resource

100%

Cache lookup attributes

100%

Cache store in outbound

100%

Token metric dimensions

100%

Token limit has counter-key

100%

Streaming incompatibility noted

100%

Backend timeout set

100%

estimate-prompt-tokens enabled

100%

Without context: $0.4936 · 2m 45s · 18 turns · 22 in / 9,302 out tokens

With context: $0.5051 · 1m 50s · 18 turns · 2,950 in / 6,534 out tokens

100%

50%

Protecting an AI Customer Service Bot from Harmful Inputs

Content safety and jailbreak protection policies

Criteria

Without context

With context

Uses llm-content-safety

100%

All four categories present

30%

100%

False-positive thresholds raised

44%

100%

On-error jailbreak handler

10%

100%

Blocked content returns 400

100%

Rate limit by agent ID

90%

100%

Rate limit includes retry header

100%

Content safety in inbound

100%

Auth policy present

37%

100%

Policy order respected

30%

100%

Threshold reasoning in notes

100%

Without context: $0.2029 · 1m 36s · 9 turns · 14 in / 4,652 out tokens

With context: $0.4424 · 1m 38s · 22 turns · 817 in / 5,111 out tokens

86%

47%

Multi-Tenant AI Gateway with Regional Load Balancing

Multi-region load balancing and multi-tenant governance

Criteria

Without context

With context

Hash-based backend selection

Retry on 429

100%

Retry cycles to next backend

100%

Per-tenant token limit

100%

Token limit value per tenant

100%

Tenant dimension in metrics

22%

100%

Managed identity auth

100%

Correct RBAC role in script

88%

Backend create commands

100%

estimate-prompt-tokens set

100%

Backend timeout elevated

100%

Policy ordering correct

33%

50%

Without context: $0.4017 · 2m 14s · 11 turns · 14 in / 9,083 out tokens

With context: $0.8018 · 3m 55s · 19 turns · 2,951 in / 13,974 out tokens

Evaluated: 2 days ago
Agent: Claude Code
Model: Unknown

Table of Contents

AI Gateway Policy Stack for Financial Services Protecting an AI Customer Service Bot from Harmful Inputs Multi-Tenant AI Gateway with Regional Load Balancing

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.