Quest: Gorse Bot 3000 — AgTech Agent
| Field | Value |
|---|---|
| Category | Creative |
| Difficulty | Master |
| Points | 200 |
| Deadline | 42 days |
| Evidence | Text + Link |
| Objectives | 7 |
Narrative Context
Gorse is the golden curse of New Zealand’s pastures—beautiful but invasive, choking out native species and productive farmland. Traditional control is expensive and labor-intensive. What if AI agents could coordinate targeted treatment, identifying gorse patches from imagery and optimizing eradication efforts? This is agentic engineering for greener pastures.
Transformation Goal
You will apply agentic AI to a real NZ agricultural problem, learning to work with domain experts, geospatial data, and practical deployment constraints.
Objectives
Objective 1: Know Your Enemy
| Field | Value | |——-|——-| | Points | 25 | | Evidence | Text + Link | | Depends On | — |
Research gorse biology, current control methods, and the scale of the problem in NZ. Interview a farmer or DOC ranger if possible.
Objective 2: Data Reconnaissance
| Field | Value | |——-|——-| | Points | 25 | | Evidence | Text + Link | | Depends On | — |
Identify available data sources: satellite imagery, drone footage, existing weed mapping projects. Document access methods.
Objective 3: Detection Design
| Field | Value | |——-|——-| | Points | 35 | | Evidence | Text + Link | | Depends On | Objectives 1, 2 |
Design an AI system for gorse detection from imagery. Could be ML classification, vision model prompting, or hybrid approach.
Objective 4: Prototype Detection
| Field | Value | |——-|——-| | Points | 45 | | Evidence | Link (code/demo) | | Depends On | Objective 3 |
Build a working gorse detector. Demonstrate on sample imagery with accuracy metrics.
Objective 5: Agent Orchestration
| Field | Value | |——-|——-| | Points | 35 | | Evidence | Text + Link | | Depends On | Objective 4 |
Design an agent system that takes detection outputs and generates treatment recommendations (location priority, method, timing).
Objective 6: Field Validation
| Field | Value | |——-|——-| | Points | 20 | | Evidence | Text | | Depends On | Objective 5 |
If possible, validate with real-world data or expert review. Document findings and next steps.
Objective 7: Open Source
| Field | Value | |——-|——-| | Points | 15 | | Evidence | Link | | Depends On | Objective 6 |
Release your code and documentation for the guild and broader community.
Resources
Design Notes
This quest brings together computer vision, agentic systems, and real-world agricultural impact—directly from the guild’s founding ideas. The open source requirement ensures community benefit.