ahacker-1/cre-agent-skills / claude-code-plugins/cre-document-ingestion
cre-document-ingestion
VerifiedCRE Document Ingestion suite — 4 specialist skills for classifying and extracting structured data from deal documents including rent rolls, T-12 financials, and offering memoranda.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 4.2 KB |
| skills/document-classifier.md | markdown | 11.1 KB |
| skills/financials-parser.md | markdown | 10.8 KB |
| skills/offering-memo-parser.md | markdown | 10.3 KB |
| skills/rent-roll-parser.md | markdown | 9.6 KB |
skills/financials-parser.md
View on GitHub---
license: Apache-2.0
metadata:
author: "Avi Hacker, J.D."
organization: "The AI Consulting Network"
homepage: https://www.theaiconsultingnetwork.com
source: https://github.com/ahacker-1/cre-agent-skills
copyright: "Copyright 2026 Avi Hacker, J.D. / The AI Consulting Network"
---
# T12 Financials Parser
Parse a trailing 12-month (T12) operating statement and extract structured revenue, expense, NOI, and per-unit metrics for underwriting analysis.
---
## When to Use This Skill
Use this skill when you have a T12 income statement, operating expense report, or trailing financial summary for a multifamily or commercial property and need to extract clean, standardized financial data from it. It handles monthly-column format, annual summary format, and budget-comparison format. Use it at the start of underwriting to establish a reliable baseline of current property performance.
---
## What You'll Need to Provide
- The T12 financial statement — as an Excel spreadsheet, CSV, or PDF
- The number of units in the property (for per-unit metric calculations)
- If the statement is a partial year (YTD) or covers a non-standard period, note that and specify the period covered
- If the document mixes actual and budgeted figures, note that so the parser can separate them
---
## Mission
Parse T12 financial statements, income/expense reports, and operating statements to extract revenue, expenses, and NOI data for underwriting.
---
## Strategy
The following sections detail the input formats recognized, header patterns used for column identification, and the step-by-step extraction logic applied to each document.
---
## Input Formats
### Common T12 Structures
**Format A: Monthly Columns**
```
Jan Feb Mar ... Dec Total
Revenue
Gross Rent 280K 282K 285K ... 295K 3.4M
Vacancy (14K) (12K) (15K) ... (10K) (170K)
...
```
**Format B: Annual Summary**
```
T12 Per Unit % of Revenue
Revenue
Gross Potential 3,600,000 18,000 100%
Vacancy (180,000) (900) (5%)
...
```
**Format C: Budget Comparison**
```
Actual Budget Variance
Revenue
Gross Rent 3,420,000 3,500,000 (80,000)
...
```
---
## Column Header Recognition
| Category | Possible Headers |
|----------|------------------|
| Revenue | Revenue, Income, Gross Potential, GPR |
| Vacancy | Vacancy, Loss, Vacancy Loss |
| Concessions | Concessions, Free Rent, Specials |
| Bad Debt | Bad Debt, Write-off, Collections |
| Other Income | Other, Other Income, Misc, Fee Income |
| Total Revenue | Total Revenue, Effective Gross, EGI, Net Rental Income |
| Expenses | Expenses, Operating Expenses, OpEx |
| Taxes | Taxes, Real Estate Taxes, Property Tax |
| Insurance | Insurance, Property Insurance |
| Utilities | Utilities, Electric, Gas, Water, Sewer |
| Repairs | Repairs, R&M, Maintenance, Repairs & Maintenance |
| Management | Management, Management Fee, Property Management |
| Payroll | Payroll, Personnel, Salaries, Wages |
| Admin | Admin, Administrative, G&A, Office |
| Marketing | Marketing, Advertising, Leasing |
| Contract | Contract Services, Vendors |
| Total Expenses | Total Expenses, Total OpEx, Operating Expenses |
| NOI | NOI, Net Operating Income, Operating Income |
---
## Extraction Logic
### Step 1: Identify Statement Structure
Determine which format by scanning:
- Monthly columns (12+ numeric columns) → Format A
- Annual summary with 1–3 numeric columns → Format B
- Columns labeled "Actual", "Budget", "Variance" → Format C
Look for "T12", "Trailing", "Annual", "YTD" indicators to confirm the period covered.
### Step 2: Find Key Sections
Locate section headers:
1. Revenue / Income section
2. Expense / Operating Expense section
3. NOI / Net Operating Income line
### Step 3: Extract Revenue Items
```json
{
"revenue": {
"grossPotentialRent": 3600000,
"vacancy": -180000,
"vacancyPercent": 0.05,
"lossToLease": -84000,
"concessions": -30000,
"badDebt": -12000,
"otherIncome": {
"petFees": 24000,
"parkingIncome": 36000,
"laundryIncome": 18000,
"lateFees": 12000,
"applicationFees": 8000,
"otherMisc": 22000,
"total": 120000
},
"effectiveGrossIncome": 3414000
}
}
```
### Step 4: Extract Expense Items
```json
{
"expenses": {
"realEstateTaxes": 320000,
"insurance": 80000,
"utilities": {
"electric": 72000,
"gas": 36000,
"water": 48000,
"trash": 24000,
"total": 180000
},
"repairsAndMaintenance": {
"general": 120000,
"turnover": 60000,
"hvac": 20000,
"total": 200000
},
"managementFee": 150000,
"managementFeePercent": 0.044,
"payroll": {
"salaries": 200000,
"benefits": 50000,
"payrollTaxes": 30000,
"total": 280000
},
"administrative": 45000,
"marketing": 35000,
"contractServices": {
"landscaping": 36000,
"pest": 12000,
"security": 24000,
"pool": 18000,
"total": 90000
},
"reserves": 60000,
"totalOperatingExpenses": 1440000
}
}
```
### Step 5: Calculate Key Metrics
```json
{
"metrics": {
"noi": 1974000,
"noiMargin": 0.578,
"expenseRatio": 0.422,
"expensePerUnit": 7200,
"revenuePerUnit": 17070,
"noiPerUnit": 9870,
"taxesPerUnit": 1600,
"insurancePerUnit": 400,
"utilitiesPerUnit": 900,
"repairsPerUnit": 1000,
"payrollPerUnit": 1400
}
}
```
### Step 6: Trend Analysis (if monthly data is available)
```json
{
"trends": {
"revenueGrowthYoY": 0.03,
"expenseGrowthYoY": 0.02,
"noiGrowthYoY": 0.04,
"occupancyTrend": "stable",
"seasonality": {
"highOccupancy": ["May", "Jun", "Jul"],
"lowOccupancy": ["Dec", "Jan", "Feb"]
}
}
}
```
---
## Output Format
Return structured JSON using this schema:
```json
{
"source": {
"file": "t12-financials.xlsx",
"period": "Jan 2025 - Dec 2025",
"extractedAt": "2026-02-01T12:00:00Z",
"confidence": 0.92
},
"summary": {
"effectiveGrossIncome": 3414000,
"totalExpenses": 1440000,
"netOperatingIncome": 1974000,
"expenseRatio": 0.422,
"noiMargin": 0.578
},
"revenue": {
"grossPotentialRent": 3600000,
"vacancy": -180000,
"vacancyPercent": 0.05,
"concessions": -30000,
"badDebt": -12000,
"otherIncome": 120000,
"effectiveGrossIncome": 3414000
},
"expenses": {
"taxes": 320000,
"insurance": 80000,
"utilities": 180000,
"repairs": 200000,
"management": 150000,
"payroll": 280000,
"admin": 45000,
"marketing": 35000,
"contract": 90000,
"reserves": 60000,
"total": 1440000
},
"perUnit": {
"units": 200,
"revenue": 17070,
"expenses": 7200,
"noi": 9870,
"taxes": 1600,
"insurance": 400
},
"warnings": [],
"dataGaps": []
}
```
---
## Quality Checks
Run these validation checks before finalizing:
1. **Math check**: Revenue − Expenses = NOI (within 1% tolerance)
2. **Expense ratio**: Use property-type context. For multifamily, 35–55% is a useful screening range; for other CRE property types, compare against sector-specific benchmarks instead of treating the multifamily range as universal
3. **Management fee**: Compare against property type and lease structure. For management-intensive assets, 3–5% of EGI is common; for leaner industrial structures it may be lower
4. **Taxes per unit**: Compare to property's known tax bill or market norms; large discrepancies may indicate the T12 is pre-reassessment
5. **Per-unit metrics**: Flag any expense category that is more than 2× or less than 0.5× typical benchmarks
6. **Below-the-line items excluded**: CapEx, debt service, and depreciation must not be included in operating expenses; identify and exclude if present
---
## When Data is Missing
### Missing Expense Categories
Some T12s combine line items differently (e.g., "Administrative & Marketing" as one line):
- Map to the closest standard category
- Note the mapping assumption in `warnings`
### Below-the-Line Items Included
CapEx, debt service, and depreciation should be excluded from operating expenses:
- Identify these items (look for "Capital Expenditures", "Depreciation", "Debt Service", "Mortgage")
- Exclude from total operating expenses
- Note in `warnings` that they were excluded and state the amount
### Partial Year Data
YTD instead of full T12:
- If more than 6 months of data: annualize by dividing total by months elapsed and multiplying by 12
- If fewer than 6 months: flag as insufficient for reliable T12 analysis
- Note the annualization basis in `warnings`
### Pro Forma Mixed In
Some statements mix actual and projected figures:
- Extract actuals only for the T12 output
- Flag any line item that appears to be a projection rather than actual (look for round numbers, footnotes about "stabilized", or budget-vs-actual formatting)
- Store pro forma items separately and note them in `warnings`
---
## Confidence Scoring
| Level | Criteria |
|-------|----------|
| **HIGH (0.90+)** | Full 12-month actual data, all major line items present, math validates |
| **MEDIUM (0.70–0.89)** | Some line items combined or missing, minor annualization needed, or one anomalous month |
| **LOW (<0.70)** | Partial year data, heavy combining of categories, pro forma mixed with actuals, or PDF extraction with ambiguous table structure |
---
## Related Knowledge Bases
For deeper analysis, pair this skill with:
- **Multifamily Benchmarks** knowledge base — to validate expense ratios, per-unit metrics, and NOI margins against market norms for the property class and geography
- **Underwriting Calculations** knowledge base — for building a forward-looking financial model from T12 baseline data
- **Rent Roll Parser** skill — to cross-validate gross potential rent and occupancy against the rent roll
- **Offering Memo Parser** skill — to cross-validate the T12 NOI against the financial highlights in the OM
---
## Attribution
Built and maintained by [The AI Consulting Network](https://www.theaiconsultingnetwork.com/?utm_source=github&utm_medium=skill-file&utm_campaign=cre-agent-skills), the commercial real estate AI consulting practice of Avi Hacker, J.D., and part of [CRE Agent Skills](https://github.com/ahacker-1/cre-agent-skills), an open-source library of AI skills for commercial real estate.
If this skill saved you time and you want systems like it built inside your firm, [reach out](https://www.theaiconsultingnetwork.com/contact?utm_source=github&utm_medium=skill-file&utm_campaign=cre-agent-skills). We would love to work with you.
Copyright 2026 Avi Hacker, J.D. / The AI Consulting Network. Licensed under the [Apache License 2.0](https://github.com/ahacker-1/cre-agent-skills/blob/main/LICENSE). This attribution notice must be retained in all copies, redistributions, and derivative works of this file.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |