Corporate Sale-Leaseback Financial Structuring Agent with Claude Agents SDK
Corporate finance directors lack real-time decision tools to structure sale-leasebacks that optimize corporate liquidity without degrading credit ratings. Build an agent system using the Claude Agents SDK to evaluate corporate property balance sheets and structure lease terms. Achieve automated deal memo generation with 95% financial accuracy.
What you are building
The core problem, expected build, and operating context for this challenge.
Develop an autonomous structuring agent using Anthropic's Claude Agents SDK to compute optimal leaseback rent, cap rate, and lease term durations.
How work is evaluated
Evaluates mathematical correctness of calculated initial rent, sale proceeds, and post-transaction FCCR.
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.
rent_calculation_check
Validates initial annual rent = property value * target cap rate within 0.1% tolerance
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
math_precision
Accuracy of financial ratio 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
Implement an agent architecture using Claude Agents SDK tool definition protocols
Calculate sale proceeds, initial cap rate, annual rental escalation, and lease duration options
Perform corporate coverage tests (Fixed Charge Coverage Ratio - FCCR post-transaction)
Output structured deal term sheets for corporate real estate decision-making
Reference links and supporting material
Corporate asset balance sheet summaries with property appraisal values and operating earnings.
Challenge setup
Evaluates mathematical correctness of calculated initial rent, sale proceeds, and post-transaction FCCR.
Challenge input
JSON containing property_appraised_value, target_cap_rate, tenant_ebitdar, annual_rent_escalation
Claude Agents SDK
Anthropic agent development framework for tool orchestration.
Evaluated output
JSON containing gross_proceeds, initial_annual_rent, post_tx_fccr, recommended_lease_years
- Validates initial annual rent = property value * target cap rate within 0.1% tolerance
- Accuracy of financial ratio calculations • target: 0.95 • range: 0-1
- Benchmark: Corporate Real Estate Structuring Benchmark
- Math Precision target: 0.95
- 1 public reference case
- Python execution harness
- Python 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
- Metaflow · Required
- Sarvam AI · Optional
- Dora · Optional
Evaluation contract
- rent_calculation_check · Weight 1
- math_precision · Weight 1
Metrics
- math_precision · Target 0.95
Benchmarks
- Corporate Real Estate Structuring Benchmark
- general
Reference cases
- 1 public reference case
Workflow tasks
- Task 1
Input: JSON containing property_appraised_value, target_cap_rate, tenant_ebitdar, annual_rent_escalation
Output: JSON containing gross_proceeds, initial_annual_rent, post_tx_fccr, recommended_lease_years
Execution harness: Python
Evaluation tools
- Claude Agents SDK · agent_framework
Recipe freeze marker
No recipe freeze time is recorded.