Multi-Agent Cyber Threat Triage
This challenge focuses on building a cutting-edge multi-agent system for autonomous cybersecurity threat triage. Participants will develop an A2A protocol-enabled agent network using LangGraph, where specialized agents collaborate to analyze, contextualize, and recommend remediation steps for simulated security alerts. The system will leverage GPT-5 Pro for advanced reasoning and threat intelligence processing, integrating with simulated Security Information and Event Management (SIEM) systems via MCP for real-time data access and automated response orchestration.
What you are building
The core problem, expected build, and operating context for this challenge.
This challenge focuses on building a cutting-edge multi-agent system for autonomous cybersecurity threat triage. Participants will develop an A2A protocol-enabled agent network using LangGraph, where specialized agents collaborate to analyze, contextualize, and recommend remediation steps for simulated security alerts. The system will leverage GPT-5 Pro for advanced reasoning and threat intelligence processing, integrating with simulated Security Information and Event Management (SIEM) systems via MCP for real-time data access and automated response orchestration.
Shared data for this challenge
Review public datasets and any private uploads tied to your build.
What you should walk away with
Master LangGraph for building stateful DAG agent workflows, including dynamic agent routing and checkpointing.
Implement the A2A protocol for secure, asynchronous agent-to-agent communication, ensuring data integrity and handoff.
Design MCP-enabled tool integration with simulated enterprise SIEM APIs for real-time alert ingestion and security orchestration (e.g., firewall rule updates).
Leverage GPT-5 Pro's extended thinking capabilities, applying adaptive reasoning budgets for deep analysis of suspicious network activities and malware signatures.
Build specialized agents within the LangGraph workflow: Alert Analyst, Threat Intelligence Researcher (RAG-enabled), Remediation Planner, and Reporting Agent.
Deploy a proof-of-concept multi-agent system using Docker for containerization and basic observability for agent interactions.
How this agent runs
Evaluation will assess the agent system's ability to accurately analyze cybersecurity alerts, propose relevant remediations, and demonstrate effective A2A communication within the LangGraph workflow.
Challenge input
{'alert_id': 'string', 'timestamp': 'string', 'source_ip': 'string', 'destination_ip': 'string', 'event_type': 'string', 'raw_log': 'string'}
Agent execution
The configured agent processes the input under the challenge policy.
Evaluated output
{'analysis': 'string', 'severity': 'low|medium|high|critical', 'threat_category': 'string', 'confidence_score': 'float'}
- Verify that the LangGraph workflow executes all defined agent nodes in the correct sequence, demonstrating...
- Confirm that critical data is passed correctly between agents using the A2A protocol (e.g., analysis from A...
- Check if the MCP integration successfully invokes simulated SIEM/SOAR APIs for remediation actions.
- AnalysisAccuracy target: 0.85
- Python execution harness
View technical recipe
Configured tools
No tool records are attached.
Evaluation contract
- The evaluation module defines the checks.
Recipe state
This is a preview. The configuration can change before the evaluation recipe is locked.
Run this agent on your dataset and AI stack
Bring your dataset, model providers, and success criteria. We will scope the right managed run for your team.
Scope a managed run[ok] Wrote CHALLENGE.md
[ok] Wrote .versalist.json
[ok] Wrote eval/examples.json
Requires VERSALIST_API_KEY. Works with any MCP-aware editor.
DocsFind another challenge
Jump to a random challenge when you want a fresh benchmark or a different problem space.