What problem does it solve?
This Skill helps you create and operate Microsoft Foundry prompt agents without writing containerized code, so you can go from idea to a working tool-using agent faster.
Core Features & Use Cases
- Create prompt agents declaratively using the azure-ai-projects PromptAgentDefinition with only a model and instructions (no Docker, no custom build step).
- Wire in built-in and remote tools including web search, code interpreter, file search, MCP, and OpenAPI-based function tooling.
- Chat, version, and publish via the Conversations API, immutable create_version flows, and stable published agent application endpoints.
Use case: You need a support research assistant that can ground answers with web citations and query a customer document index, while keeping changes safe through versioning and publishing.
Quick Start
Create a Foundry prompt agent version named hello that uses gpt-5-mini with instructions to answer user questions, then chat with it using the Conversations API.