microsoft-foundry

Microsoft Foundry agent deployment involves creating hosted or prompt agents, configuring tools, deploying images, and verifying runtime behavior. This Skill orchestrates invocation testing and trace analysis to ensure reproducible end-to-end workflows.

Yes, you can manage RBAC, permissions, connections, quota, and regional capacity for Foundry projects. This Skill handles infrastructure operations to set up resources and maintain access controls across your AI delivery environments.

The best way to run evaluations is to execute evaluation runs, analyze traces, and curate datasets. This Skill manages the quality and iteration loops including fine-tuning models and maintaining local metadata for reproducible workflows.

Yes, you need Azure CLI or Foundry MCP context resolution to manage Foundry agents. This Skill requires these environments along with local metadata overlays and cache-aware artifact handling to execute reproducible end-to-end workflows.

To troubleshoot a failed Microsoft Foundry agent invocation, you inspect failures using trace analysis and then improve prompts or training data based on evaluation results. This Skill enables continuous improvement through local metadata and cache artifacts.