Skip to main content
Back to challenges
Challenge

MCP Server for Enterprise Sustainability Reporting

Develop a cutting-edge agent system designed for enterprise sustainability reporting. This system will leverage the MCP to securely and interpretably access diverse, simulated internal data sources (e.g., IoT sensor readings from factories, energy consumption logs, supply chain data). Agents, orchestrated in a stateful graph, will analyze this data using Claude Opus 4.5 for high-level reasoning and GPT-4o for tool orchestration, integrating a custom anomaly detection model served by TorchServe. The goal is to generate dynamic, accurate sustainability and compliance reports, identify environmental risks, and provide actionable recommendations for optimization.

Workflow AutomationHosted by Vera
Sign in free to start
Challenge brief

What you are building

The core problem, expected build, and operating context for this challenge.

Develop LangGraph agents leveraging MCP for secure enterprise data access, using Claude Opus 4.1 and GPT-4o, and integrating a TorchServe-hosted model, to generate comprehensive sustainability reports.

Delivery guide

How work is evaluated

Evaluation

The evaluation will focus on the LangGraph system's ability to accurately access and interpret simulated enterprise data via MCP, identify sustainability risks, ensure compliance, and generate high-quality, actionable reports. Test cases will cover diverse data scenarios and policy checks.

Datasets

Challenge datasets

Public datasets, plus the private datasets you have access to.

Learning goals

What you should walk away with

  • Master LangGraph for building stateful, cyclic agent workflows, enabling complex multi-step reasoning, dynamic tool calling, and explicit state management for enterprise-grade applications.

  • Implement the MCP for secure and auditable access to simulated enterprise data sources, ensuring data governance and privacy are maintained throughout the agent workflow.

  • Orchestrate a multi-LLM strategy: Utilize Claude Opus 4.1 for high-level analytical reasoning, complex report generation, and nuanced policy interpretation, while employing OpenAI GPT-4o for efficient tool orchestration, data extraction, and summarization.

  • Deploy a custom data analysis model (e.g., an anomaly detection model for unusual energy spikes) using TorchServe, and integrate this served model as a specialized tool callable by LangGraph agents.

  • Design and implement distinct LangGraph nodes for data ingestion (via MCP), sustainability analysis, compliance checking, risk assessment, and report generation, ensuring seamless transitions between agent responsibilities.

  • Develop robust evaluation harnesses to verify the accuracy of data interpretation, the correctness of compliance checks, the quality of risk identification, and the clarity and actionability of generated sustainability reports.

  • Implement a continuous feedback loop and monitoring strategy for the LangGraph system, allowing for iterative improvements to agent reasoning, tool integration, and reporting accuracy based on expert review.

Resources and assets

Reference links and supporting material

Dataset notes

Sample data for 1 tasks

Challenge setup

The evaluation will focus on the LangGraph system's ability to accurately access and interpret simulated enterprise data via MCP, identify sustainability risks, ensure compliance, and generate high-quality, actionable...

Challenge input

{'data_access_request': {'mcp_token': 'string', 'data_sources': ['list of strings']}, 'sustainability_policies': ['list of strings'], 'reporting_pe...

Agent execution

No agent tool is listed for this challenge.

Evaluated output

{'report_summary': 'string', 'resource_consumption_analysis': {'water': 'float', 'energy': 'float', 'raw_materials': 'float'}, 'identified_risks':...

Evaluation checks
  • Confirm that the 'mcp_access_log' shows successful and authorized access to simulated data sources based on...
  • Verify if the 'compliance_status' (PASS/FAIL) correctly reflects the analysis of 'identified_risks' against...
Configured evaluation evidence
  • Risk Identification Accuracy target: 90
Configured runtime
  • Python execution harness

Summary shows the first task, up to three tools, and up to three evidence items. Open the technical recipe to view all configured items.

View technical recipe

Configured tools

No tool records are attached.

Evaluation contract

  • Confirm that the 'mcp_access_log' shows successful and authorized access to simulated data sources based on the provided token.
  • Verify if the 'compliance_status' (PASS/FAIL) correctly reflects the analysis of 'identified_risks' against 'sustainability_policies'.

Metrics

  • Risk Identification Accuracy · Target 90
  • Recommendation Quality · Target 4

Workflow tasks

  1. Task 1

    Input: {'data_access_request': {'mcp_token': 'string', 'data_sources': ['list of strings']}, 'sustainability_policies': ['list of strings'], 'reporting_period': 'string'}

    Output: {'report_summary': 'string', 'resource_consumption_analysis': {'water': 'float', 'energy': 'float', 'raw_materials': 'float'}, 'identified_risks': ['list of strings'], 'recommendations': ['list of strings'], 'compliance_status': 'PASS|FAIL', 'mcp_access_log': ['list of strings']}

    Execution harness: Python

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about MCP Server for Enterprise Sustainability Reporting