leestott

leestott / foundry-cicd

Public

Enterprise-ready CI/CD reference for Microsoft Foundry AI agents, with parallel GitHub Actions and Azure DevOps pipelines, evaluation-driven quality gates, and controlled promotion across Dev, Test, and Production.

agent-evaluation ai-agents azure azure-devops cicd
12
3
89% credibility
Found May 29, 2026 at 15 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
AI Summary

A reference implementation of automated pipelines that build, test, evaluate, and deploy AI agents across development, testing, and production environments using either GitHub Actions or Azure DevOps.

How It Works

1
📋 Discover the reference pipelines

You find ready-made templates that show exactly how to automatically build, test, and release AI agents across different environments.

2
Choose your automation platform
🐙
GitHub Actions

Use GitHub's built-in automation to run your pipelines whenever code changes

☁️
Azure DevOps

Use Azure's professional-grade pipelines with built-in approval workflows

3
🔍 Your code gets automatically checked

Every time you share new code, the system automatically checks for problems, runs tests, and scans for security issues.

4
✅ Quality gates verify your AI behaves correctly

Before your AI can move forward, it must prove it answers correctly, completes tasks reliably, and responds quickly enough.

5
🧪 Deploy to a testing environment

Your AI agent is deployed to a safe development space where you can try it out without affecting real users.

6
🚀 Promote through stages with team approval

With approval from your team, your AI progresses through increasingly important stages until it's ready for real users.

🎉 Your AI agent is live and working

Your AI agent is now live, serving real users, with automated checks keeping everything running smoothly.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 15 to 12 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is foundry-cicd?

This is a reference implementation for deploying AI agents on Microsoft Foundry using enterprise CI/CD practices. It provides two ready-to-use pipeline definitions: one for GitHub Actions and one for Azure DevOps. The pipelines handle the full lifecycle from code commit through validation, evaluation, and promotion across Dev, Test, and Production environments. Quality gates enforce measurable standards like hallucination rates, task completion thresholds, and latency targets before agents can advance between stages.

Why is it gaining traction?

The AI agent deployment space lacks established patterns for continuous delivery. This repository fills that gap by combining static analysis, security scanning, unit testing, and model evaluation into a single controlled pipeline. The evaluation-driven gates are the differentiator--rather than just checking if code runs, these pipelines verify that agents actually perform correctly using defined datasets. Having parallel implementations for both GitHub Actions and Azure DevOps makes it accessible regardless of your existing infrastructure.

Who should use this?

Teams building AI agents on Microsoft Foundry who need a proven deployment process. Platform engineers establishing CI/CD standards for AI workloads will find the multi-environment promotion model useful. Organizations requiring audit trails and approval gates before production deployments will appreciate the structured approach. If you are starting with Foundry and want to avoid reinventing pipeline infrastructure, this gives you a solid foundation to customize.

Verdict

The concept is solid and the execution matches real-world enterprise needs. However, the project has only 12 stars and a credibility score under 1%, indicating it is early and unproven at scale. The documentation is clear and governance files are present, but test coverage and community feedback remain unknowns. Use this as a learning reference or starting template, but validate thoroughly before committing to production workloads.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.