Skip to content

bipul049/cloudops-intelligence-hub

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

🖥️ CloudOps Intelligence Hub

An AI-powered AWS operations assistant built with Claude Code and a custom MCP (Model Context Protocol) server.

What It Does

Natural language queries against live AWS infrastructure — no dashboards, no scripts, just ask Claude.

Architecture

You (natural language) ↓ Claude Code (terminal agent) ↓ Custom MCP Server (Python/boto3) ↓ AWS Services (EC2, S3, CloudWatch, Cost Explorer) ↓ Markdown Reports & Runbooks (auto-saved)

Features

  • 🔍 EC2 Health — list instances, state, AZ, CPU utilisation
  • 🪣 S3 Insights — bucket inventory, object count, size
  • 🚨 CloudWatch Alarms — real-time alarm state monitoring
  • 💰 Cost Explorer — month-to-date spend by service
  • 📋 Ops Digest — single command full environment report
  • 🔥 Incident Runbooks — auto-generated and saved to /runbooks

Slash Commands

Command Description
/ops-digest Full AWS environment report saved to /reports
/runbook-high-cpu <instance-id> Incident runbook for high CPU

Tech Stack

  • Claude Code — agentic terminal runtime
  • MCP (Model Context Protocol) — tool interface
  • Python 3.11 + boto3 — custom MCP server
  • AWS Free Tier — EC2, S3, CloudWatch, Cost Explorer
  • GitHub MCP — automated runbook commits

Project Structure

cloudops-hub/ ├── CLAUDE.md # Persistent agent context ├── mcp-server/ │ └── server.py # Custom MCP server (6 AWS tools) ├── .claude/ │ ├── commands/ # Slash command definitions │ └── skills/ # Compound workflow skills ├── reports/ # Auto-generated ops digests └── runbooks/ # Auto-generated incident docs

Setup

# 1. Clone the repo
git clone https://github.com/<yourname>/cloudops-intelligence-hub
cd cloudops-intelligence-hub

# 2. Create Python environment
conda create -n cloudops python=3.11 -y
conda activate cloudops
pip install "mcp[cli]" boto3

# 3. Configure AWS credentials
aws configure --profile cloudops-mcp

# 4. Register MCP server with Claude Code
claude mcp add cloudops-mcp -- python3 mcp-server/server.py

# 5. Launch
claude

IAM Permissions Required

AmazonEC2ReadOnlyAccess AmazonS3ReadOnlyAccess CloudWatchReadOnlyAccess AWSBillingReadOnlyAccess

Example Output

Ask Claude Code:

  • "List all my EC2 instances and their CPU utilisation"
  • "What have I spent on AWS this month by service?"
  • "Are there any CloudWatch alarms in ALARM state?"
  • "Generate a full ops digest report"

CCA Exam Domains Covered

Domain Coverage
Agentic Architecture & Orchestration (27%) ✅ Multi-tool orchestration via MCP
Tool Design & MCP Integration (18%) ✅ Custom MCP server with 6 tools
Claude Code Configuration (20%) ✅ CLAUDE.md, skills, slash commands
Prompt Engineering & Structured Output (20%) ✅ JSON schemas, structured responses
Context Management & Reliability (15%) ✅ Persistent context via CLAUDE.md

Built by Bipul Singh — AWS Cloud Architect

About

AI-powered AWS ops assistant built with Claude Code + custom MCP server

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages