Challenge

Geo-Compliance Satellite Swarm

Inspired by Planet Labs' recent decision to withhold satellite imagery of conflict regions at the request of the US government, you will design a multi-agent 'Compliance Swarm'. Using CrewAI, you will define three distinct roles: a Geo-Strategist, a Legal Compliance Officer, and an Image Analyst. The swarm will automate the process of reviewing satellite image requests against dynamic international conflict zone databases and government directives. The system will use GPT-5.4 Pro for high-level synthesis and strategic reasoning, while Claude Sonnet 4.6.6 handles the detailed regulatory text analysis. You will host the specialized image classification models on the Hugging Face Inference API and use Text Generation Inference (TGI) for high-speed local processing of the agentic communication. The agents must collaborate to decide if an image of a specific coordinate can be released, ensuring that all military and law enforcement anti-drone operations.

Workflow AutomationHosted by Vera
Challenge brief

What you are building

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

Inspired by Planet Labs' recent decision to withhold satellite imagery of conflict regions at the request of the US government, you will design a multi-agent 'Compliance Swarm'. Using CrewAI, you will define three distinct roles: a Geo-Strategist, a Legal Compliance Officer, and an Image Analyst. The swarm will automate the process of reviewing satellite image requests against dynamic international conflict zone databases and government directives. The system will use GPT-5.4 Pro for high-level synthesis and strategic reasoning, while Claude Sonnet 4.6.6 handles the detailed regulatory text analysis. You will host the specialized image classification models on the Hugging Face Inference API and use Text Generation Inference (TGI) for high-speed local processing of the agentic communication. The agents must collaborate to decide if an image of a specific coordinate can be released, ensuring that all military and law enforcement anti-drone operations.

Datasets

Shared data for this challenge

Review public datasets and any private uploads tied to your build.

Loading datasets...
Evaluation rubric

How submissions are scored

These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.

Dimensions
2 scoring checks
Binary
2 pass or fail dimensions
Ordinal
0 scaled dimensions
Dimension 1

Zero-Release Integrity

System must never approve imagery within 50km of restricted zones.

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Dimension 2

Consensus Time

Time taken for all 3 agents to reach a verdict • target: 10 • range: 1-30

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Learning goals

What you should walk away with

  • Master CrewAI role definitions, goals, and backstories for collaborative agent tasks

  • Implement inter-agent delegation and hierarchical process flows in CrewAI

  • Integrate the Hugging Face Inference API for triggering computer vision tasks within an LLM workflow

  • Optimize model serving using Text Generation Inference (TGI) for internal agent dialogue

  • Orchestrate a multi-model approach using GPT-5.4 Pro for strategy and Claude Sonnet 4.6.6 for legal verification

  • Build a compliance-first toolset that restricts agent access based on geographic fence-lines

How this agent runs

The swarm is tested by providing coordinates within a known conflict zone and a request from a civilian entity.

Preview configuration

Challenge input

Latitude, Longitude, Requesting Entity Type

CrewAI

Framework for orchestrating

GPT-5

Policy Serving in the agent workflow.

Evaluated output

Approval Status, Risk Level, Agent Consensus log

Checks for
  • System must never approve imagery within 50km of restricted zones.
  • Time taken for all 3 agents to reach a verdict • target: 10 • range: 1-30
Proof of success
  • Consensus Time target: 10
  • 1 public reference case
Runtime evidence
  • Docker execution harness
View technical recipe

Configured tools

Action Space
  • CrewAI · Required
  • crewAI · Optional
Policy Serving
  • GPT-5 · Optional

Evaluation contract

  • Zero-Release Integrity · Weight 1
  • Consensus Time · Weight 1

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
Start from your terminal
$npx -y @versalist/cli start geo-compliance-satellite-swarm

[ok] Wrote CHALLENGE.md

[ok] Wrote .versalist.json

[ok] Wrote eval/examples.json

Requires VERSALIST_API_KEY. Works with any MCP-aware editor.

Docs
Manage API keys
Explore

Find another challenge

Jump to a random challenge when you want a fresh benchmark or a different problem space.

Useful when you want to pressure-test your workflow on a new dataset, new constraints, or a new evaluation rubric.

Frequently Asked Questions about Geo-Compliance Satellite Swarm