V

vertex-ai

A flexible AI coding agent for the CLI, powered by a customizable prompt engine.

Atlas Architect: Your AI Co-pilot for Secure Cloud Infrastructure

This project is an AI-powered DevSecOps agent that lives within GitLab. It proactively analyzes Infrastructure-as-Code (IaC) files, specifically Terraform, to automatically visualize, secure, and optimize a developer's Google Cloud architecture before it's ever deployed.

When a developer submits a Merge Request with Terraform changes, a CI/CD pipeline triggers the agent to post a detailed analysis back as a comment. This provides instant visibility and governance, helping teams build better, safer cloud infrastructure, faster.

Key Features:

AI-Powered Visualization: Generates architecture diagrams from Terraform code using Google's Vertex AI. Security & Cost Analysis: Identifies security vulnerabilities and cost inefficiencies based on best practices. Intelligent Remediation: Automatically suggests code changes to fix identified issues. Vector-Powered Knowledge Base: Uses a MongoDB Atlas Vector Search index of official Google Cloud documentation to provide highly relevant, context-aware explanations for its recommendations.

Core Technologies:

Platform: GitLab CI/CD, Google Cloud Platform (GCP), MongoDB Atlas Services: Google Cloud Run, Google Cloud Build, Google Vertex AI, MongoDB Atlas Vector Search Frameworks & Languages: Python, Flask, Gunicorn

Terminal Chat Completion client for Google's Gemini AI models written in Go

AI-powered security scanner that finds vulnerabilities and provides one-click fixes directly in GitLab merge requests. A reusable CI/CD Catalog component built with Google Cloud Vertex AI.

Euclidean Distance or Cosine Similarity using CSV files as vectors embedding DBs - written in Go