<div align="center">

# 🤖 Agentic AI & Model Context Protocol (MCP) DevOps Platform

[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
[![Python 3.10+](https://img.shields.io/badge/python-3.10+-blue.svg)](https://www.python.org/downloads/)
[![Model Context Protocol](https://img.shields.io/badge/Protocol-Model_Context_Protocol_(MCP)-00C7B7.svg)](https://modelcontextprotocol.io/)
[![Multi-Cloud](https://img.shields.io/badge/Multi--Cloud-Azure_•_AWS_•_GCP-0078D4.svg)](https://github.com/satyajeetkumarsing)
[![Tests: Passing](https://img.shields.io/badge/tests-6%20passed-brightgreen.svg)](https://github.com/satyajeetkumarsing/agentic-ai-mcp-platform)

**An autonomous multi-agent platform connecting Frontier LLM reasoning cores (Claude 3.5 Sonnet, GPT-4, Gemini) to enterprise multi-cloud control planes via the Model Context Protocol (MCP).**

Designed & Engineered by **[Satyajeet Kumar Singh](https://github.com/satyajeetkumarsing)**

</div>

---

## 🧭 Overview & Architectural Vision

Traditional DevOps automation relies on static scripts, brittle regex parsers, and manual human gating. The **Agentic AI MCP DevOps Platform** bridges the cognitive capabilities of Large Language Models with production cloud infrastructure:

```text
                                  +---------------------------------------+
                                  |     Frontier LLM Reasoning Engine     |
                                  |  (Claude 3.5 Sonnet • GPT-4o • Gemini)|
                                  +-------------------+-------------------+
                                                      │
                                          Standard JSON-RPC 2.0
                                      Model Context Protocol (MCP)
                                                      │
                         +----------------------------+----------------------------+
                         │                                                         │
                         ▼                                                         ▼
         +-------------------------------+                         +-------------------------------+
         |    Multi-Cloud Ops Tools      |                         |   IaC & Terraform MCP Server  |
         +-------------------------------+                         +-------------------------------+
         | • query_cloud_resources       |                         | • inspect_tfplan              |
         | • triage_incident (OOM/504)   |                         | • detect_state_drift          |
         | • triage_security_vulnerabilities                       | • calculate_infracost         |
         +-------------------------------+                         +-------------------------------+
                         │                                                         │
                         +----------------------------+----------------------------+
                                                      │
                                                      ▼
                                       +-------------------------------+
                                       |   Autonomous Agent Workflows  |
                                       +-------------------------------+
                                       | • PR Reviewer Agent           |
                                       | • Drift Remediation Agent     |
                                       | • FinOps Cost Optimizer Agent |
                                       +-------------------------------+
```

---

## 🛠️ MCP Toolset Specification

This platform implements the standard JSON-RPC 2.0 Model Context Protocol (MCP) tool contract:

| Tool Name | Parameters | Description |
|:---|:---|:---|
| `inspect_tfplan` | `plan_json`, `enforce_strict_security` | Scans Terraform plans for CIS benchmark violations and flags destructive deletion operations. |
| `detect_state_drift` | `resource_address`, `live_state`, `declared_state` | Performs attribute-level comparison and synthesizes corrective HCL blocks. |
| `calculate_infracost` | `resource_changes`, `currency` | Computes monthly infrastructure financial delta before pull requests are merged. |
| `triage_incident` | `service_name`, `log_sample`, `environment` | Analyzes container crash logs (e.g. OOMKilled, 504 timeouts) and prescribes self-healing actions. |
| `query_cloud_resources` | `cloud_provider`, `query_filter`, `scope` | Discovers multi-cloud resources across Azure Resource Graph, AWS Resource Explorer, and GCP Asset Inventory. |

---

## 🤖 Built-In Autonomous Agents

### 1. Shift-Left PR Review Agent (`src/agents/pr_review_agent.py`)
- Automatically triggered on Pull Request open/update.
- Concurrently runs IaC security inspection and Infracost financial estimation.
- Comments directly with human-readable Markdown verdicts and automated approvals.

### 2. Autonomous Drift Healer (`src/agents/drift_remediation_agent.py`)
- Continuously audits live cloud infrastructure against GitOps state.
- Automatically generates dedicated branch, commits corrective HCL, and opens a Pull Request.

### 3. FinOps Optimization Auditor (`src/agents/finops_optimizer_agent.py`)
- Identifies unattached managed disks, orphaned public IPs, and non-prod instances running off-hours.
- Calculates projected annualized savings (~28%-35% average reduction).

---

## 🚀 Quickstart & Usage

### 1. Clone & Run Tests
```bash
git clone https://github.com/satyajeetkumarsing/agentic-ai-mcp-platform.git
cd agentic-ai-mcp-platform

# Run all unit tests
python3 -m unittest discover tests
```

### 2. Run Autonomous Demos
```bash
# Shift-left PR security & cost review
python3 -m src.cli demo-pr-review

# Autonomous drift detection & HCL generation
python3 -m src.cli demo-drift

# FinOps multi-cloud waste optimization
python3 -m src.cli demo-finops

# Incident telemetry & OOM triage
python3 -m src.cli demo-incident
```

### 3. Claude Desktop / Cursor MCP Setup
Add this server to your Claude Desktop config (`claude_desktop_config.json`):

```json
{
  "mcpServers": {
    "agentic-ai-devops": {
      "command": "python3",
      "args": ["-m", "src.server"],
      "cwd": "/path/to/agentic-ai-mcp-platform",
      "env": {
        "PYTHONPATH": "."
      }
    }
  }
}
```

---

## 🐳 Docker Deployment

```bash
docker build -t satyajeetkumarsing/agentic-ai-mcp-platform:latest .
docker run -i --rm satyajeetkumarsing/agentic-ai-mcp-platform:latest serve
```

---

## 📄 License
This project is licensed under the MIT License - see the LICENSE file for details.

Developed with precision by **[Satyajeet Kumar Singh](https://github.com/satyajeetkumarsing)**.