skills.rest
davidrrowley
davidrrowleyCOMMUNITY
Explore all skills by this author →
Website𝕏@davidrrowleyGitHub

microsoft-foundry

Manage Azure AI Foundry agent lifecycles from deployment to evaluation.

Updated Apr 12, 2026
Search Tags:#deployment#azure#evaluation#rbac#agents#quota#foundry
One-click install
npx skills add https://github.com/davidrrowley/CortexYouV3 --skill microsoft-foundry-davidrrowley
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: microsoft-foundry
Source: https://github.com/davidrrowley/CortexYouV3/tree/main/.agents/skills/microsoft-foundry
Command: npx skills add https://github.com/davidrrowley/CortexYouV3 --skill microsoft-foundry-davidrrowley

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Managing the full lifecycle of Azure AI Foundry agents—project onboarding, model discovery and deployment, container builds, ACR publishing, runtime invocation, evaluation, tracing, and remediation—is complex and error-prone without a coordinated, opinionated workflow. Teams struggle with quota and region selection, RBAC and identity configuration, reliable containerization and ACR workflows, and trace-to-dataset evaluation loops that tie production failures back to reproducible test cases. This Skill centralizes guardrails, pre-checks, and repeatable workflows so teams can deploy and maintain agents safely and consistently.

Core Features & Use Cases

  • Intelligent model and capacity discovery across regions and projects with handoffs to quick (preset) or fully customized deployment flows.
  • Hosted-agent container lifecycle: Dockerfile generation guidance, ACR build/push patterns, agent creation, container start/stop, status polling, and verification.
  • Prompt and hosted agent operations: create prompt agents, invoke multi-turn sessions, vNext session handling, and test invocations.
  • Eval-driven optimization: harvest production traces into versioned datasets, auto-create evaluators (two-phase strategy), run batch evals, cluster failures, and drive prompt optimization → redeploy → re-evaluate loops.
  • Operational tooling: quota and PTU management, RBAC/managed identity guidance, App Insights KQL templates for trace analysis, and troubleshooting container logs and telemetry correlation.
  • Use cases: onboard a new Foundry project, deploy a hosted agent to ACR, run batch evaluations from traces, detect regressions, and set up CI/CD evaluation pipelines.

Quick Start

Deploy an agent named support-agent to Foundry using the existing ACR image contosoregistry.azurecr.io/support-agent:20240601.

Frequently Asked Questions about microsoft-foundry

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I deploy an Azure AI Foundry agent using an existing ACR image?▼

To deploy an Azure AI Foundry agent, you reference an existing ACR image like `contosoregistry.azurecr.io/support-agent:20240601` during the hosted-agent creation flow. The Skill handles agent creation, container start, status polling, and deployment verification.

Can I run batch evaluations on Azure AI Foundry agents using production traces?▼

Yes, you can harvest production traces into versioned datasets and run batch evaluations on Azure AI Foundry agents. The workflow auto-creates evaluators, clusters failures, and drives prompt optimization loops to detect regressions before redeployment.

What is the process for onboarding a new project and discovering models in Azure AI Foundry?▼

Project onboarding in Azure AI Foundry involves intelligent model discovery across regions and projects, validating quota and RBAC configurations, and setting up managed identities. This ensures safe, coordinated deployment environments before agent creation begins.

Does this workflow support App Insights trace correlation for troubleshooting Foundry agents?▼

Yes, the workflow provides App Insights KQL templates for trace analysis and telemetry correlation. It enables troubleshooting container logs and tracing production failures back to reproducible test cases for Azure AI Foundry agents.

How do I manage quota and PTU allocations when deploying Foundry agents across multiple regions?▼

Managing quota and PTU allocations across regions requires intelligent capacity discovery and pre-checks before deployment. The Skill validates region availability, applies guardrails, and ensures quota limits are checked during the Azure AI Foundry agent lifecycle.

What's the best way to automate the container lifecycle for hosted Foundry agents?▼

The best way to automate the hosted-agent container lifecycle is using guided Dockerfile generation, ACR build/push patterns, and automated container start/stop with status polling. This ensures reliable containerization and consistent ACR publishing workflows.

Related Skills

Explore More Software Engineering Skills→
paidethonpaidethon

microsoft-foundry

Manage Azure AI Foundry agent and model lifecycles, including deployment and evaluation.

Community
Advanced
microsoftmicrosoft
242

azure-kusto

Run KQL queries and analyze data in Azure Data Explorer.

Official
Advanced
p-potvinp-potvin
1

microsoft-foundry

Deploy and manage Azure AI Foundry agents across container build, ACR push, hosting, and monitoring.

Community
Advanced
duylinhdang1998duylinhdang1998

ml-infrastructure

Build and deploy ML training pipelines with Docker, Kubernetes, and FastAPI.

Community
Advanced