Skip to main content
Back to challenges
Challenge

Develop a Secure, A2A Protocol-Enabled GovTech Agent

This challenge focuses on creating a secure, compliance-aware AI agent system designed for sensitive government applications. You will develop a multi-agent system leveraging the A2A Protocol for secure, verifiable agent-to-agent communication and MCP-enabled tool integration for interacting with simulated government APIs and data sources. The core of the challenge is to use GPT-5.1 for advanced reasoning and Semantic Kernel for orchestrating agents that can handle sensitive data, adhere to strict regulatory compliance, and operate across a simulated multi-cloud environment (e.g., AWS and Azure government clouds). The agents should demonstrate extended thinking with adaptive reasoning budgets to perform complex, secure data analysis or policy enforcement tasks, ensuring data integrity and access control through A2A and MCP standards.

Special Purpose AgentsHosted 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 a secure, multi-agent GovTech system using GPT-5.1 and Semantic Kernel, featuring A2A Protocol communication and MCP tool integration.

Delivery guide

How work is evaluated

Evaluation

The GovTech agent system will be evaluated on its ability to securely process a sensitive government data request, demonstrating strict adherence to A2A communication, Model Context Protocol (MCP) tool usage, and compliance rules. Success criteria include data integrity, access control, and the accuracy of the final output.

Datasets

Challenge datasets

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

Learning goals

What you should walk away with

  • Utilize extended thinking with adaptive reasoning budgets for complex, sensitive tasks.

  • Design Model Context Protocol (Model Context Protocol)-enabled tools for secure interaction with simulated government APIs.

  • Orchestrate multi-agent workflows for GovTech applications using Semantic Kernel with GPT-5.

  • Implement A2A Protocol for secure, verifiable agent-to-agent communication.

  • Integrate compliance and security constraints into agent decision-making processes.

Resources and assets

Reference links and supporting material

Dataset notes

Sample data for 1 tasks

Challenge setup

The GovTech agent system will be evaluated on its ability to securely process a sensitive government data request, demonstrating strict adherence to A2A communication, Model Context Protocol (MCP) tool usage, and comp...

Challenge input

{'request_id': 'REQ-2025-001', 'requester_id': 'Agency_X', 'data_subject_id': 'Citizen_A', 'data_elements': ['full_name', 'address', 'medical_histo...

Agent execution

No agent tool is listed for this challenge.

Evaluated output

{'request_id': 'REQ-2025-001', 'status': 'completed', 'output_data': {'full_name': 'Citizen A', 'address': '****', 'medical_history': '****'}, 'aud...

Evaluation checks
  • All specified policy rules (PII redaction, access control) are correctly enforced, as indicated by 'complia...
  • At least 3 valid A2A protocol signatures are present in the 'audit_log', demonstrating inter-agent communic...
Configured evaluation evidence
  • RedactionAccuracy target: 95
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

  • All specified policy rules (PII redaction, access control) are correctly enforced, as indicated by 'compliance_report'.
  • At least 3 valid A2A protocol signatures are present in the 'audit_log', demonstrating inter-agent communication.

Metrics

  • RedactionAccuracy · Target 95
  • ReasoningEfficiency · Target 15

Workflow tasks

  1. Task 1

    Input: {'request_id': 'REQ-2025-001', 'requester_id': 'Agency_X', 'data_subject_id': 'Citizen_A', 'data_elements': ['full_name', 'address', 'medical_history'], 'policy_rules': {'PII_redaction_required': true, 'access_control': ['Agency_X']}}

    Output: {'request_id': 'REQ-2025-001', 'status': 'completed', 'output_data': {'full_name': 'Citizen A', 'address': '****', 'medical_history': '****'}, 'audit_log': [{'agent': 'PolicyAgent', 'action': 'Verified_Access', 'result': 'success', 'a2a_signature': '...'}], 'compliance_report': {'PII_redaction': 'compliant'}}

    Execution harness: Python

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about Develop a Secure, A2A Protocol-Enabled GovTech Agent