Build Urban Adaptive Reuse Valuation Pipeline with Mastra AI & Claude Code
London's King's Cross demonstrated how transforming historic industrial assets into mixed-use urban spaces yields massive long-term asset appreciation. Replicating this success requires analyzing complex zoning rules, heritage preservation codes, and historical build constraints. In this challenge, developers build an adaptive reuse intelligence pipeline using Mastra AI and Claude Code CLI tools. The system ingests architectural CAD outlines, historical land-use records, and zoning bylaws to generate optimal adaptive reuse recommendations (e.g., converting industrial heritage structures into retail/office hubs) along with estimated ROI.
What you are building
The core problem, expected build, and operating context for this challenge.
Develop an adaptive reuse valuation agent using Mastra AI and Claude Code to evaluate industrial heritage property transformations.
How work is evaluated
Evaluates the agent's accuracy in recommending optimal adaptive reuse options and calculating projected internal rate of return (IRR).
Challenge datasets
Public datasets, plus the private datasets you have access to.
How submissions are scored
These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.
IRR Target Range Check
Ensures projected IRR falls within realistic range (12% - 18%)
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
Feasibility Metric Match
Accuracy of financial and conversion cost calculations • target: 0.95 • range: 0-1
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
What you should walk away with
Configure Mastra AI TypeScript workflows to parse geospatial and heritage planning constraints
Utilize Claude Code automation scripts for generating dynamic architectural feasibility summaries
Model adaptive reuse conversion costs against projected yield metrics (IRR, Cap Rate)
Produce automated urban regeneration report packages for municipal approval
Reference links and supporting material
Dataset of 25 heritage industrial property blueprints, historical preservation restrictions, and London market demand tables.
Challenge setup
Evaluates the agent's accuracy in recommending optimal adaptive reuse options and calculating projected internal rate of return (IRR).
Challenge input
JSON with site_sqft, historic_grade, acquisition_cost, local_demand_sector
Mastra AI
Required framework for TypeScript workflow orchestration.
Claude Code
Required tool for dynamic code-driven execution.
Evaluated output
JSON with recommended_use, total_conversion_cost, projected_cap_rate, projected_irr_pct
- Ensures projected IRR falls within realistic range (12% - 18%)
- Accuracy of financial and conversion cost calculations • target: 0.95 • range: 0-1
- Benchmark: UrbanReuseBench 2025
- Feasibility Metric Match target: 0.95
- 1 public reference case
- JavaScript execution harness
- javascript sandbox (unavailable on Versalist)
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
- Mastra AI · Required
- Claude Code · Optional
- Capably · Optional
- Claude Code · Optional
- Mastra AI · Required
Evaluation contract
- IRR Target Range Check · Weight 1
- Feasibility Metric Match · Weight 1
Metrics
- Feasibility Metric Match · Target 0.95
Benchmarks
- UrbanReuseBench 2025
- general
Reference cases
- 1 public reference case
Workflow tasks
- Task 1
Input: JSON with site_sqft, historic_grade, acquisition_cost, local_demand_sector
Output: JSON with recommended_use, total_conversion_cost, projected_cap_rate, projected_irr_pct
Execution harness: JavaScript
Evaluation tools
- Mastra AI · Agent Framework
- Claude Code · AI Engineering Tooling
Recipe freeze marker
No recipe freeze time is recorded.