Home / Projects / lra-ai-platform

Active Development

lra-ai-platformCore Platform

lra-ai-platform is a multi-agent orchestration platform for DevOps automation. It implements a deterministic task-planning architecture in which a Supervisor decomposes natural language intents into executable plans, routes them through a Governance Engine enforcing five RBAC permission levels, and delegates execution to eight specialized agents — Cloud Architect, DevOps Engineer, Security Engineer, SRE, Documentation, Reviewer, OpenShift and Founder. Each agent invokes real integrations: Terraform, Kubernetes, Ansible, AWS, Azure, GCP, GitHub, Trivy and Checkov, among others.

Python 3.11+ FastAPI boto3 Kubernetes Terraform Ansible Trivy Checkov

View on GitHub Request demo

Agents

Integrations

Workflows

Tests · 0.14s

What it does

8 agents. 14 integrations. One governance engine.

lra-ai-platform routes intents to specialized agents via keyword matching. Each agent executes real tasks through live integrations — no mocks, no simulations. Every action passes through a 5-level RBAC engine before execution.

lra scan — security audit on any GitHub repository

infrastructure_review — multi-cloud AWS/Azure/GCP analysis

security_review — Trivy + Checkov on every push

pr_review — automated code review on pull requests

documentation_update — README, ADR, Runbooks from code

deploy_eks — full EKS deployment pipeline

REST API at localhost:8000 — 13+ endpoints

CLI: lra init · lra scan · lra review · lra run · lra status

# lra-ai-platform CLI

$ lra scan --github lra-cloud-ops/aws-terraform-devops

→ Scanning repository...

→ Running Trivy + Checkov...

✓ Critical: 0

✓ Checkov: 47/50 passed

⚠ Medium: 2 → documented

$ lra review aws

✓ EKS: healthy · 3 nodes

✓ RDS: running · 99.9%

✓ ArgoCD: synced · 0 drift

✓ Completed in 4.2s

Agents

8 specialized agents

Founder Agent

Creates GitHub repos, generates README, ARCHITECTURE, ROADMAP

github

Cloud Architect

Reviews AWS, Azure, GCP infrastructure and Well-Architected compliance

aws azure gcp github

DevOps Engineer

Executes Terraform, kubectl, Helm, Ansible operations

terraform kubernetes ansible

Security Engineer

Runs Trivy and Checkov scans on images, IaC, K8s, GitHub Actions

trivy checkov github

SRE Agent

Configures Prometheus alerts, Grafana dashboards, CloudWatch alarms

prometheus grafana cloudwatch

OpenShift Agent

Manages OpenShift projects, deployments, operators via oc CLI

openshift kubernetes

Documentation Agent

Generates README, ADR, Runbooks from code analysis

github local

Reviewer Agent

Reviews PRs, analyzes code quality, posts GitHub comments

github trivy checkov

Governance

5-level RBAC. Fail-safe by default.

Every task passes through the Governance Engine before execution. Deny by default.

Level 1

READ_ONLY

List repos, read metrics

Level 2

PROPOSE

Create PRs, generate docs

Level 3

DEVELOPMENT

Commits, deploy to dev

Level 4

PRODUCTION

Deploy to prod

Level 5

ADMIN

Modify agents and policies

Explore the source code

All code is public. Clone it, run it, evaluate it.

View on GitHub → Get in touch

Ecosystem

Role in the platform

lra-ai-platform functions as the AI orchestration layer of the LRA CloudOps platform. Agents execute security scans against aws-terraform-devops via SecurityAgent, generate and maintain documentation across all repositories via DocumentationAgent, and monitor cluster health on k8s-on-premise via SREAgent.

Scans security of aws-terraform-devops →

SecurityAgent runs Trivy + Checkov on Terraform modules and K8s manifests.

Reviews and documents all projects →

ReviewerAgent and DocumentationAgent run across all ecosystem repositories.

Provisions infrastructure on aws-terraform-devops →

CloudArchitectAgent can trigger Terraform plan/apply workflows.

Monitors platform on k8s-on-premise →

SREAgent checks cluster health, resource usage and GitOps sync status.