Turn free A.CRE models into a Claude underwriting skill.
Download professional commercial real estate models, let Claude study how they work, and turn the methodology into one reusable CRE underwriting workflow.
Open the A.CRE Model Hub and download the commercial real estate Excel models you want Claude to learn from. Keep the original files intact so Claude can inspect the formulas, tabs, assumptions, and outputs.
Open Claude and start a new conversation. Upload the downloaded Excel models. If the library is too large for one upload, work in logical batches (multifamily, development, office, retail, etc.) and tell Claude that every batch belongs to the same skill-building project.
03
Paste the prompt below
Use the prompt below to tell Claude to reverse-engineer the models rather than merely summarize them. It instructs Claude to preserve the underwriting logic, build model-selection rules, separate facts from assumptions, and package the result as a reusable skill.
04
Have Claude build the skill
Let Claude inspect the models and propose the skill architecture. Review the model inventory and methodology it identifies. If something important is missing, add the relevant model or correct the assumption before Claude finalizes the skill files.
05
Save the skill in Claude
Once Claude has created the skill, save/install it using Claude’s Skills interface available on your plan. The exact UI can change, so follow Claude’s current Skills flow if the button wording differs. Then test it on a real or sample deal before relying on the output.
Step 3
Paste this into Claude.
Upload the models first, then paste this prompt. Claude should analyze the actual files you provide; the prompt does not magically contain the A.CRE models itself.
Copy + paste
CRE underwriting skill prompt
I’m giving you a collection of commercial real estate Excel models from A.CRE (Adventures in CRE).
I want you to turn these files into a reusable Commercial Real Estate Underwriting Skill for Claude.
Do not simply summarize the spreadsheets. Reverse-engineer how the models work, understand the formulas and underwriting logic, and build a skill that can use those methods on future commercial real estate deals.
First, analyze every model I uploaded and identify:
- property type and deal type
- required inputs and assumptions
- major formulas and calculations
- cash-flow structure
- financing and debt calculations
- exit assumptions
- return metrics such as IRR, equity multiple, cash-on-cash return, cap rate, yield on cost, DSCR, debt yield, LTV and LTC where relevant
- sensitivity analyses
- waterfalls/promotes where relevant
- any model-specific methodology worth preserving
Then identify the common underwriting framework across the files and create property-specific workflows for the property types represented in the models, such as multifamily, office, retail, industrial, hotel, self-storage, mixed-use, development, value-add and land when applicable.
Build model-selection logic so that when I give Claude a deal, it first determines:
1. property type
2. transaction type
3. information provided
4. information missing
5. which uploaded A.CRE methodology is the best fit
6. what analysis can reasonably be completed with the available information
The skill should accept deal materials such as offering memorandums, rent rolls, T-12s, operating statements, Excel files, loan terms, broker emails, PDFs and my own assumptions.
For every deal, clearly separate:
- facts from source documents
- assumptions
- calculated outputs
Never invent missing numbers or silently fill missing assumptions. Ask me for important missing information or clearly label any assumption required to proceed.
The skill should be able to:
- underwrite acquisitions, developments and value-add deals where supported by the uploaded models
- analyze revenue, vacancy, expenses and NOI
- analyze debt and financing
- calculate relevant levered and unlevered returns
- run appropriate sensitivities
- identify key risks and return drivers
- calculate useful breakpoints such as maximum purchase price for a target IRR when possible
- produce a concise investment conclusion with the reasoning behind it
For each completed underwriting, structure the output as:
1. Deal Snapshot
2. Key Assumptions
3. Underwriting Results
4. What I Like
5. What Concerns Me
6. Sensitivity Analysis
7. Breakpoints
8. Investment Conclusion
Create the actual reusable Claude Skill with a focused SKILL.md plus supporting reference files where useful. Keep large reference material outside SKILL.md so the skill remains efficient. Preserve useful methodology from the uploaded A.CRE models rather than reinventing it.
Before finalizing the skill, show me:
1. which uploaded models you analyzed
2. what each model is designed to do
3. the common framework you extracted
4. the property-specific methodologies you identified
5. the skill architecture you propose
Then build and package the skill so I can save it and reuse it in future Claude conversations.
After the skill is built, give me 3 test prompts I can use to verify that it correctly recognizes and underwrites a commercial real estate deal.
Test it before trusting it
Your first underwriting test
Give the finished skill an offering memorandum, rent roll, and T-12 for a sample deal and ask: “Underwrite this deal. Separate source facts, assumptions, and calculated outputs. Tell me what information is missing before making unsupported assumptions.” Then compare the output against the original model.
This workflow is for analysis and education. Verify calculations and assumptions before making investment decisions.
Commercial construction
Want to see how I’m implementing AI inside commercial construction companies?
I’m hosting a free commercial construction AI masterclass covering where AI actually fits, what to build first, and what to ignore.