Autonomous AI Sustainability Intelligence Platform

EcoSphere AI is a full-stack multi-agent sustainability intelligence system that analyzes environmental challenges and generates practical, data-driven climate solutions.

Built using an ADK Multi-Agent Architecture and MCP Server Design, the platform enables users to describe sustainability challenges and receive structured advisory reports including environmental analysis, recommended actions, implementation timelines, impact estimation, and knowledge-backed insights.

Designed for the Kaggle AI Agent Intensive Vibe Coding Capstone — Freestyle Track, demonstrating a production-style local-first agent workflow.

• AI Sustainability Advisory System

  • Converts environmental challenges into actionable sustainability strategies.

• ADK Multi-Agent Collaboration

  • Includes specialized agents working together:

    • Environmental Problem Analyzer Agent
    • Research Agent
    • Solution Planner Agent
    • Impact Assessment Agent
    • Final Advisor Agent

• MCP Server Architecture

  • Provides controlled access to:

    • Sustainability knowledge retrieval
    • Environmental analysis tools
    • Impact calculation utilities

• Local-First AI Workflow

  • Runs with a simulated offline environment without dependency on external APIs.

• Premium Sustainability Dashboard

  • Modern dark eco-tech interface built using React, Vite, and custom CSS.

• Interactive Sustainability Report

  • Generates:

    • Problem analysis
    • Research insights
    • Action roadmap
    • Environmental impact summary
    • References

• Impact Intelligence

  • Calculates sustainability indicators:

    • Carbon reduction
    • Water conservation
    • Waste reduction
    • Estimated benefits

• Secure Agent Execution

  • Includes:

    • Input validation
    • Safe tool execution
    • Error handling
    • Injection protection

The system uses a sequential agent pipeline where each agent improves the previous context.

graph TD

User([Environmental Challenge])

User --> Analyzer[Environmental Problem Analyzer Agent]

Analyzer --> Researcher[Research Agent]

Researcher --> Planner[Solution Planner Agent]

Planner --> Assessor[Impact Assessment Agent]

Assessor --> Advisor[Final Advisor Agent]

Advisor --> Dashboard([EcoSphere AI Dashboard])

EcoSphere AI uses an MCP-style tool layer to separate agent reasoning from utility operations.

Agents communicate with tools through:

executeMCPTool(toolName, parameters)

Available tools:

• Validates user input
• Detects unsafe patterns
• Prevents malicious execution

• Searches local sustainability knowledge base
• Retrieves relevant environmental information

• Calculates estimated:

  • Carbon impact
  • Water savings
  • Waste reduction
  • Financial benefits

EcoSphere_AI/

├── backend/
│   ├── src/
│   │   ├── agents/
│   │   │   ├── analyzerAgent.js
│   │   │   ├── researchAgent.js
│   │   │   ├── plannerAgent.js
│   │   │   ├── impactAgent.js
│   │   │   └── advisorAgent.js
│   │   ├── mcp/
│   │   │   ├── mcpServer.js
│   │   │   └── tools/
│   │   │       ├── knowledgeBase.js
│   │   │       ├── calculator.js
│   │   │       └── validator.js
│   │   └── server.js
│   ├── data/
│   │   └── sustainability_db.json
│   └── package.json
├── frontend/
│   ├── src/
│   │   ├── components/
│   │   │   ├── ProblemInput.jsx
│   │   │   ├── StatusTracker.jsx
│   │   │   ├── ReportViewer.jsx
│   │   │   ├── TimelineView.jsx
│   │   │   └── ImpactSummary.jsx
│   │   ├── App.jsx
│   │   └── index.css
│   └── package.json
└── README.md

Requirements:

• Node.js v18+

cd backend

npm install

npm start

Backend runs:

http://localhost:5000

cd frontend

npm install

npm run dev

Frontend runs:

http://localhost:3000

Example:

User:

"How can my university reduce food waste?"

Process:

1. Analyzer Agent

   • Understands the environmental issue

2. Research Agent

   • Retrieves sustainability knowledge

3. Planner Agent

   • Creates implementation roadmap

4. Impact Agent

   • Estimates possible environmental benefits

5. Advisor Agent

   • Generates final sustainability intelligence report

Output:

• Problem Analysis
• Recommended Actions
• Timeline
• Impact Summary
• References

EcoSphere AI includes:

• Input sanitization
• Length validation
• Unsafe content filtering

• Controlled MCP tool access
• Isolated execution flow

• Agent failure recovery
• Safe fallback responses
• Frontend error protection

The system verifies:

• Agent workflow completion
• Input validation
• MCP tool execution
• Knowledge retrieval
• Impact calculation accuracy

Run:

npm test

• PDF sustainability report export
• Excel impact data export
• Sustainability score generation
• SDG goal mapping
• Personalized follow-up AI advisor
• Historical report storage

EcoSphere AI demonstrates how autonomous AI agents can transform environmental knowledge into practical sustainability decisions.

By combining:

• Multi-Agent AI
• MCP Tool Architecture
• Secure execution
• Environmental intelligence

the platform acts as an AI sustainability consultant for real-world climate challenges.