Imported from DataSift-Ty-Personal/SiftStack (skills/rehab-estimator/SKILL.md). Install upstream with npx skills add DataSift-Ty-Personal/SiftStack --skill rehab-estimator. Copyright stays with the author.

Rehab Cost Estimator

Produces professional, itemized renovation cost estimates for investment properties. Two exit strategies analyzed side-by-side: Full Rehab (investor flip to retail) and Wholetail (light cosmetic for quick sale to another investor or retail-lite buyer).

All pricing is calibrated to the local market where the property sits — not national averages.

Single-Estimate Approach

This skill produces ONE cost estimate per line item — not a Low/Mid/High range. The estimate targets the low-to-mid range of real-world contractor pricing (roughly the 35th percentile between Low and Mid reference prices in rehab-categories.md). This keeps estimates conservative and closer to what actual contractor bids come back at, based on real deal calibration data showing that Low/Mid/High ranges led to over-estimation. When in doubt, lean toward the lower end — real contractors tend to come in closer to low than mid on most line items.

When to Read Reference Files

This skill uses progressive disclosure. Read these files at the specific steps indicated:

Workflow Overview

Step 1: Gather Property Intel
Step 2: Photo & Condition Analysis
Step 2.5: Read Calibration Data (MANDATORY — read real-deal-calibration.md before scoping)
Step 3: Build Scope of Work (read rehab-categories.md, finish-tiers.md, wholetail-vs-rehab.md)
Step 4: Localize Pricing (read local-pricing-guide.md)
Step 5: Cross-Reference Comps (validate against Bucket B finishes)
Step 5.5: Calculate Full Deal Costs (financing, holding, buying, selling)
Step 6: Generate Deliverables (run scripts)
Step 7: Present to User

Step 1: Gather Property Intel

Collect these inputs from the user. Some may come from uploaded files (comp reports, inspection reports, MLS listings), photos, or conversation:

Required:

• Property address (full: street, city, state, zip)
• Square footage (GLA)
• Bed/bath count
• Year built
• Current condition description OR photos

Helpful but optional:

• Comp report or ARV (from the comping skill or user-provided)
• Purchase price / contract price
• Specific known issues (roof age, HVAC status, foundation, etc.)
• User's target exit strategy preference
• User's budget constraints
• Whether the user has a GC or is self-managing
• Financing details (loan type, LTV, interest rate, points) — defaults to hard money assumptions if not provided
• Hold time estimate (months from purchase to sale) — will be dynamically calculated from rehab scope + comp DOM if not provided
• Comp report days-on-market (DOM) — if the comping skill was run, pull median DOM from Bucket B renovated comps to estimate resale timeline
• Property tax amount (annual) — will research if not provided
• Whether property has HOA fees

If the user provides a comp report (especially one from the real-estate-comping skill), extract:

• Bucket A median PPSF (unrenovated)
• Bucket B median PPSF (renovated)
• Renovation premium percentage
• What finishes the renovated comps had (if noted in comp comments)
• Subject property condition notes

Step 2: Photo & Condition Analysis

If the user provides property photos or an inspection report, analyze them systematically. Go room by room, system by system:

Exterior Checklist

Interior Checklist

For each item, assign a condition rating:

• Good — Functional and presentable, no work needed
• Fair — Functional but dated, cosmetic update recommended for flip
• Poor — Needs replacement or significant repair
• Missing/Failed — Non-functional or absent, must address

Step 2.5: Read Calibration Data (MANDATORY)

Read references/real-deal-calibration.md NOW — before building any scope of work.

This file contains verified unit prices from a real contractor SOW and the scoping rules that prevent over-estimation. Key rules:

Scoping Discipline — Room-by-Room, NOT System-by-System

Real contractors scope by walking room to room and noting ONLY what each room needs. The skill must do the same:

1. Walk each room from the inspection photos
2. For each room, note ONLY what visibly needs to change
3. Common per-room items: paint, flooring, outlets/switches ($100/room), light fixtures, doors
4. NOT common: new windows (only if damaged/failed), trim (only if damaged)

DO Scope:

• Items visibly damaged, broken, or missing
• Items outdated enough to hurt resale (old cabinets, dated fixtures when comps show modern)
• Code requirements (smoke detectors, GFCI outlets in wet areas)
• Cosmetic updates that comps show (LVP flooring, updated lighting)
• Appliances — ALWAYS include a full appliance package in the full rehab scope when renovated comps show updated appliances (almost always). This is the most commonly missed category. Typical SS package: fridge ($1,400) + dishwasher ($550) + microwave ($300) + stove ($700) = $2,950. Typical W/B package: fridge ($650) + dishwasher ($450) + microwave ($250) + stove ($450) = $1,800. Include even if existing appliances "work" — dated appliances kill buyer perception in a flip.
• AC unit — include when the existing system is >15 years old OR when the unit looks visibly dated/worn in photos, even if technically functional. Budget $2,800 for AC replacement. Buyers and inspectors flag old HVAC systems. Exception: if confirmed replaced within 2-3 years, exclude it.
• Smoke detectors — always include in any full rehab. Budget $30/ea × (bedrooms + 1 per floor). Code requirement.

DON'T Scope:

• HVAC if confirmed newer (≤3 years old) by seller, listing, or visible manufacturer date — note "recently replaced, pending inspection"
• Full panel upgrade unless evidence of undersized (fuse box, < 100 amp)
• System-wide outlet replacement — only replace in rooms being renovated
• All baseboard replacement — only where visibly damaged or missing. But DO add shoe mold wherever new flooring is installed.
• Staging, photography, permits — these are NOT contractor costs (separate budget items)
• Roof — do NOT include roof replacement in the base estimate unless the roof is actively failing (visible holes, tarps, severe structural sag). Instead, note as "⚠️ Needs inspection — budget $7,000-$12,000 separately if replacement required." The roof is the single largest line item that inflates estimates when included speculatively. Real investors verify roof condition before committing this cost. If the user specifically asks to include the roof OR provides evidence of active failure, include it.
• Every window — only replace damaged, single-pane, or failed-seal windows. Count from photos, don't assume all need replacing. When photo evidence is ambiguous, estimate conservatively — err on the LOWER count and note uncertainty.
• Garage door — do NOT include garage door replacement ($1,000-$1,700) unless the door is visibly broken (panels missing, severe damage, won't open). A dated-looking garage door alone is not sufficient reason to replace. Note as optional: "⚠️ Garage door replacement if needed: $1,000-$1,700."
• Speculative exterior items — do NOT add tuckpointing, gutter repair, or other exterior items unless they are CLEARLY visible in photos as damaged or failing. Age alone is not enough to include these. Only scope soffit/fascia patches where visible damage is confirmed.

Demo & Cleanup Scoping

Do NOT default to "full house demo + 2 dumpsters" unless the property genuinely requires gut-level demolition. Most investor flips need targeted demo, not a full gut:

• Light cleanup = general cleanup + hauling ($1,000-$1,500) — for houses needing cosmetic updating but not trashed
• Moderate demo = targeted demo (kitchen cabinets, bath tile, old flooring) + 1 dumpster ($2,000-$3,000) — most common for standard flips
• Full demo = gut demo + 2 dumpsters ($4,000-$5,000) — ONLY for hoarder houses, fire damage, or true gut-to-studs Ask: "How much am I ripping out?" If the answer is "kitchen cabinets, bathroom tile, and old carpet" — that's moderate demo at most.

The 60-65% Sanity Check

After building the estimate, check $/SF against these benchmarks:

• $25-35/SF = Standard investor flip (not a gut) in a Tier 2 market
• $40-55/SF = Heavy renovation with systems (plumbing, electrical, HVAC)
• $60-80/SF = Full gut-to-studs rebuild only

If your estimate exceeds $40/SF on a standard flip, you are almost certainly over-scoping. Go back and ask: "Would a real contractor actually replace this, or would they leave it alone?"

Paint Estimation Rule

Quote paint as a SINGLE all-in rate (walls + trim + ceilings combined):

• Calculate: GLA × 3.5 = total paintable wall SF
• Apply rate: $2.00-$3.00/SF (Tier 2) or $3.00-$4.50/SF (Tier 1/HCOL)
• Add primer at HALF the total SF × $2.00/SF only if needed (dark colors, stains, smoke)
• Do NOT break paint into 3 separate line items (walls, trim, ceiling)

Step 3: Build Scope of Work

Read references/rehab-categories.md now — it contains every line item with calibrated pricing. Read references/finish-tiers.md now — it defines material quality levels. Read references/wholetail-vs-rehab.md now — it defines scope differences.

Build TWO parallel scopes of work:

Full Rehab Scope

Target: Retail-ready, matches or exceeds renovated comps (Bucket B). The goal is to produce a home that competes with the best renovated comps in the market. Think "HGTV ready."

Finish level: Tier 3: Mid-Range Upgrade (see finish-tiers.md) unless comp analysis suggests otherwise.

Include every item rated Fair or worse. Also include items rated Good that are visually dated even if functional (e.g., brass fixtures in a 2025 market, oak cabinets when comps show white shaker).

Wholetail Scope

Target: Clean, functional, move-in ready but NOT retail-show-ready. Buyer is often another investor or a budget-conscious retail buyer. The goal is maximum ROI with minimum spend — fix what's broken, clean what's dirty, and don't over-improve.

Finish level: Tier 1: Minimum Viable to Tier 2: Builder Grade (see finish-tiers.md).

Only include items rated Poor or Missing/Failed, plus:

• Deep clean throughout
• Fresh paint (walls and trim)
• Basic flooring if carpet is stained/worn
• Any safety/code issues
• Curb appeal basics (pressure wash, basic landscaping)

Pricing Hierarchy — Repair Cheat Sheet First

The company's Repair Cheat Sheet is the standard estimating tool. It defines how the team budgets deals. The AI MUST produce estimates that align with the cheat sheet's methodology and pricing — NOT the contractor SOW line-item detail. The contractor SOW (in real-deal-calibration.md) is verification data showing that cheat sheet prices are in the right ballpark, but the cheat sheet is the budgeting standard.

Pricing source priority (highest to lowest):

1. User-provided pricing — always overrides everything
2. LOCKED MASTER MATERIAL LIST (data/master_material_list_37914.csv): the MATERIAL source for Knoxville / Knox County / Blount County properties. These are exact, locked Home Depot SKU prices pulled at delivery zip 37914; the team has committed to this list, so for the vast majority of material selections pull from it and nothing else. Grade map: Tier 1 = budget_price, Tier 2 = standard_price, Tier 3 = upgrade_price, Tier 4 = off-list (premium/custom is quote territory). Rows with kind = allowance are the standard non-SKU numbers (dumpster pulls, termite treatment, quartz per-sf, scope kits). Off-list exception: an item the 88-row list does not carry (an outstanding random issue, like a specialty door or a pool part) prices from the cheat sheet below; say so in the item's notes. The locked list covers MATERIALS only: labor always comes from the cheat sheet / calibration data.
3. Repair Cheat Sheet prices: the company's estimating standard (see table below). The blended labor+materials model for ALL non-Knox markets, for LABOR everywhere, and for off-list items.
4. real-deal-calibration.md verified prices — contractor verification data (use only when cheat sheet has no price for an item)
5. rehab-categories.md reference prices — national reference fallback (adjust by local cost index)

Complete Repair Cheat Sheet (Tier 2 / Dayton, OH baseline):

All prices include labor + materials. These are the company's standard estimating prices, verified against real contractor SOWs.

Flooring:

Tile vs LVP Decision Rules:

• Tile ($12/SF): Bathrooms (always), entryways, kitchens (only if ARV > $250K and comps show tile)
• LVP ($3.50/SF): All main living areas, bedrooms, hallways, secondary spaces
• Premium LVP ($5.50/SF): Only when ARV > $300K AND Bucket B comps specifically show premium brands (Lifeproof, COREtec)

Paint:

GLA Measurement for Paint Estimation:

• GLA = Gross Living Area = above-grade finished living space
• Excludes: unfinished basement, garage, covered porches, attic
• Paint formula: GLA x 3.5 = approximate paintable wall surface area
• Includes: closets, hallways, all finished rooms
• Example: 1,900 SF GLA x 3.5 = 6,650 SF paintable area x $3.00/SF = $19,950

Walls & Drywall:

Kitchen:

Bathrooms:

Electrical:

Doors & Windows:

Trim:

HVAC & Plumbing:

AC Unit Selection Rules:

• $2,500: straight swap, same tonnage, existing ductwork and pad are fine
• $3,500: replacement requiring minor ductwork modification or pad relocation
• $4,500: new install where none existed, or significant capacity upgrade (e.g., 2-ton to 3.5-ton)
• Default to $2,800 when details are unknown and system is >15 years old
• Exclude ONLY if confirmed replaced within 2-3 years

Appliances:

Appliance Package Selection Criteria:

• Stainless Steel (SS) package ($2,950): ARV > $200K, Bucket B comps show stainless, standard for most flips
• White/Black (W/B) package ($1,800): ARV < $200K, rental-grade finishes, budget flip where comps show W/B
• ALWAYS include an appliance package for full rehab -- dated appliances kill buyer perception even if functional

Exterior:

Demo & Cleanup:

Granularity rule: When the cheat sheet uses a lump sum (e.g., "Light fixtures: $1,000"), estimate as a lump sum — do NOT itemize individual fixtures. When the cheat sheet uses per-unit pricing (e.g., "Windows: $550/EA"), estimate per-unit with a conservative quantity. Match the cheat sheet's level of detail.

Quantity Estimation

For each line item, estimate quantities based on:

• Square footage — flooring, paint, drywall (use GLA + waste factor)
• Room counts — bathrooms, bedrooms drive fixture counts
• Linear feet — trim, baseboard (estimate from room perimeters: typical 12x12 room = 48 LF minus doors/windows)
• Unit counts — doors, windows, outlets, fixtures (estimate from bed/bath count and typical home layout)
• Photos — use visible evidence to refine estimates

Typical estimation shortcuts when exact measurements aren't available:

CRITICAL: Complete Room Inventory

The most common scoping error is forgetting transitional and secondary spaces. After scoping all main rooms, do a second pass:

Rule: Every space a buyer walks through must look finished. If it has walls and a floor, it needs to be in the scope.

Small Items — Covered by Contingency

Do NOT add a separate "small items" category. Items like caulk, registers, underlayment, storm doors, ice maker kits, and cabinet pulls are covered by the contingency percentage (15-20%). Itemizing these separately on top of contingency double-counts them and inflates the estimate by $400-700.

The only exception: if the user specifically asks for a fully itemized SOW with every small item broken out, then include them — but reduce the contingency to 10% to avoid double-counting.

Step 4: Localize Pricing

Read references/local-pricing-guide.md now.

National averages are a starting point, not the answer. Every market has different labor rates and material availability. The local-pricing-guide contains the methodology for adjusting costs based on:

1. Geographic cost index — use RSMeans or similar regional multipliers
2. Metro vs rural — labor availability affects rates
3. State/local permit costs — varies dramatically
4. Market-specific material costs — lumber varies by region, concrete by climate zone
5. Seasonal factors — winter construction in cold climates costs more

Use web search to find current labor rates for the specific city/metro area:

• Search: "{city} {state}" contractor labor rates {current_year}
• Search: "{city} {state}" renovation cost per square foot {current_year}
• Search: "{metro area}" home renovation costs {current_year}
• Search: "cost of living index" "{city}" OR "{metro area}"

Apply the local adjustment multiplier to every line item's labor component. Materials adjust less dramatically but still vary by region (10-20% swing in lumber/concrete).

Knox-market exception: NEVER multiply locked material prices. Prices in data/master_material_list_37914.csv are already local to zip 37914 (Home Depot delivery-zip pricing). Applying the 0.88/0.86 regional multiplier to them double-discounts materials by ~12%. The multiplier applies only to labor and to blended cheat-sheet lines; locked material prices are used as-is.

Step 5: Cross-Reference Against Comps

This step ensures the rehab budget makes economic sense. If the user provided comp data (especially from the real-estate-comping skill):

Validate the Spread

Renovation Spread = Bucket B Median PPSF - Bucket A Median PPSF
Market Premium = Spread × Subject GLA = What the market pays for renovation

Full Rehab Budget should be 40-70% of Market Premium
  → If budget > 70% of premium: warn — thin margin, risk of over-improvement
  → If budget < 40% of premium: great margin, but verify scope isn't too light

Match Comp Finishes

Look at what the renovated comps (Bucket B) actually had:

• If comps had granite counters, budget granite (not laminate)
• If comps had LVP flooring, budget LVP (not carpet or tile)
• If comps had stainless appliances, budget stainless
• The rehab should MATCH the renovated comps, not exceed them — over-improvement kills ROI

Wholetail Validation

Wholetail Target Price = Bucket A Median PPSF × Subject GLA + 10-15% (light premium for clean/functional)
Wholetail Spread = Wholetail Target - Purchase Price - Wholetail Rehab Cost

Step 5.5: Calculate Full Deal Costs

A rehab estimate without the full deal cost picture is dangerously misleading. The renovation cost is often less than half the total cost of doing the deal. Financing, holding, and transaction costs can easily add $25K-40K+ to the total, and ignoring them makes profit projections look much better than reality.

The skill must calculate ALL deal costs — not just ARV minus purchase minus rehab. Real investors use a full deal analyzer that accounts for every dollar in and out.

When to Calculate

Calculate full deal costs whenever the user provides (or you can determine) both:

• A purchase price or contract price
• An ARV or estimated sale price

If either is missing, still produce the rehab estimate but note in caveats that deal analysis requires purchase price and ARV.

Financing Costs

If the user provides specific financing terms, use those. Otherwise, apply these hard money defaults (the most common financing for flips):

Hold Time Calculation (Dynamic):

Instead of a static 4-month default, calculate hold time from the scope of work and comp data:

Total Hold Time = Rehab Duration + Marketing/Sale Period

Rehab Duration = estimated from scope complexity (see table below)
Marketing/Sale Period = median DOM from Bucket B renovated comps (if available)
                        OR default 45-60 days if no comp DOM available

Rehab Duration Benchmarks (by scope):

Round up to the nearest month for calculation. Add 2-4 weeks buffer for permitting delays in markets that require permits for rehab work.

Marketing/Sale Period:

• If the comping skill was run, pull median DOM from the Bucket B (renovated) comps. This is the most accurate predictor of how long a renovated property takes to sell in that specific market.
• If DOM data is available: Marketing/Sale Period = Bucket B median DOM (convert to months by dividing by 30, round up)
• If DOM data is NOT available: Default to 1.5 months (45 days) for Tier 2 markets, 1 month for hot markets, 2 months for slow markets
• Add 30 days for closing process on top of DOM (buyer financing, inspections, etc.)

Example (dynamic calculation):

• Full cosmetic rehab → 2 months rehab duration
• Bucket B median DOM = 28 days → ~1 month marketing + 1 month closing = 2 months sale period
• Total hold time = 4 months

Example (no comp DOM available):

• Full rehab, no structural → 3 months rehab duration
• No DOM data → default 1.5 months marketing + 1 month closing = 2.5 months → round to 3 months
• Total hold time = 6 months

If the user provides a specific hold time, always use that instead of calculating.

Financing cost formula:

Loan Amount = Purchase Price + Rehab Cost
Monthly Interest = Loan Amount × (Annual Rate / 12)
Total Financing Cost = Monthly Interest × Hold Time (months)

Example: $202,000 loan × (12% / 12) = $2,020/month × 4 months = $8,080

Holding Costs (Monthly)

Research or estimate these based on the property's location:

Total holding = Monthly total × Hold time (months)

Hold time here uses the dynamically calculated total (rehab duration + marketing/sale period) from the financing section above. This means holding costs scale accurately with the actual project timeline rather than a static assumption.

For a Tier 2 market (Midwest), typical monthly holding is $700-$1,000/month.

Buying Transaction Costs

Note: Title insurance uses state-regulated tiered rate schedules — not a flat percentage. The ~0.77% rate is calibrated from real Ohio closings and is a reasonable default for Tier 2 markets. Rates vary by state.

Selling Transaction Costs

True Profit Calculation

Net Profit = ARV
           - Purchase Price
           - Rehab Cost
           - Financing Costs
           - Holding Costs
           - Buying Costs
           - Selling Costs

True ROI = Net Profit / (Purchase Price + Rehab Cost + Financing + Holding + Buying)

This is the number that matters. A deal that looks like $46K profit on simple math might actually be $31K after all costs — that's a 32% reduction. Getting this wrong means making bad buying decisions.

Step 6: Generate Deliverables

Compile all analysis into a JSON data structure, then run both report generators.

Data Structure

{
  "property": {
    "address": "string",
    "city": "string",
    "state": "string",
    "zip": "string",
    "county": "string",
    "gla": 0,
    "beds": 0,
    "baths": 0.0,
    "year_built": 0,
    "lot_size": "string",
    "property_type": "string",
    "condition_summary": "string"
  },
  "comp_reference": {
    "arv": 0,
    "bucket_a_ppsf": 0.0,
    "bucket_b_ppsf": 0.0,
    "renovation_premium_pct": 0.0,
    "market_premium_dollars": 0,
    "comp_finish_notes": "string",
    "purchase_price": 0
  },
  "condition_assessment": [
    {
      "category": "string",
      "item": "string",
      "condition": "Good|Fair|Poor|Missing",
      "notes": "string",
      "photo_reference": "string"
    }
  ],
  "local_pricing": {
    "city": "string",
    "state": "string",
    "cost_index": 0.0,
    "labor_rate_multiplier": 0.0,
    "material_rate_multiplier": 0.0,
    "notes": "string",
    "sources": ["string"]
  },
  "rehab_estimate": {
    "scope": "Full Rehab",
    "finish_tier": "Investor-Flip Grade",
    "categories": [
      {
        "category": "string",
        "items": [
          {
            "item": "string",
            "include": true,
            "qty": 0,
            "unit": "string",
            "unit_cost": 0.0,
            "total": 0.0,
            "notes": "string"
          }
        ],
        "category_total": 0.0
      }
    ],
    "subtotal": 0,
    "contingency_pct": 0.10,
    "contingency": 0,
    "grand_total": 0,
    "cost_per_sf": 0.0
  },
  "wholetail_estimate": {
    "scope": "Wholetail",
    "finish_tier": "Builder to Mid Grade",
    "categories": [
      "...same structure as rehab_estimate categories..."
    ],
    "subtotal": 0,
    "contingency_pct": 0.05,
    "contingency": 0,
    "grand_total": 0,
    "cost_per_sf": 0.0
  },
  "deal_analysis": {
    "arv": 0,
    "purchase_price": 0,
    "rehab_cost": 0,
    "wholetail_cost": 0,
    "hold_time_months": 4,
    "hold_time_breakdown": {
      "rehab_duration_weeks": 0,
      "rehab_duration_months": 0,
      "comp_dom_days": null,
      "marketing_period_months": 0,
      "closing_period_months": 1,
      "source": "calculated|user_provided",
      "notes": ""
    },
    "financing": {
      "loan_amount": 0,
      "ltv_pct": 1.0,
      "interest_rate_annual": 0.12,
      "points_pct": 0,
      "monthly_interest": 0,
      "total_financing_cost": 0,
      "notes": "Hard money, 100% LTV on purchase + rehab"
    },
    "holding_costs": {
      "property_taxes_monthly": 0,
      "insurance_monthly": 0,
      "hoa_monthly": 0,
      "gas_monthly": 0,
      "water_monthly": 0,
      "electric_monthly": 0,
      "other_monthly": 0,
      "total_monthly": 0,
      "total_holding": 0
    },
    "buying_costs": {
      "escrow_attorney": 900,
      "title_insurance": 0,
      "title_insurance_pct": 0.0077,
      "total_buying": 0
    },
    "selling_costs": {
      "realtor_pct": 0.05,
      "realtor_fees": 0,
      "title_insurance": 0,
      "title_insurance_pct": 0.0064,
      "escrow_attorney": 900,
      "recording_fees": 400,
      "transfer_tax": 0,
      "transfer_tax_pct": 0.0012,
      "home_warranty": 750,
      "total_selling": 0
    },
    "rehab_total_all_in": 0,
    "wholetail_total_all_in": 0,
    "rehab_net_profit": 0,
    "wholetail_net_profit": 0,
    "rehab_roi_pct": 0.0,
    "wholetail_roi_pct": 0.0,
    "mao_75_pct": 0,
    "rehab_margin_vs_premium": 0.0
  },
  "material_specs": [
    {
      "category": "string",
      "item": "string",
      "specification": "string",
      "est_price": 0.0,
      "finish": "Budget|Standard|Upgrade",
      "sku": "string",
      "alt_option": "string",
      "notes": "string"
    }
  ],
  "recommendations": ["string"],
  "caveats": ["string"]
}

material_specs (populates the Material Specs sheet): For Knoxville / Knox / Blount properties, build one row per in-scope material from data/master_material_list_37914.csv. est_price = the locked price at the grade matching the finish tier (Tier 1 = budget, Tier 2 = standard, Tier 3 = upgrade); finish = that grade name; sku = the product title plus its Home Depot link from the CSV; alt_option = the approved Budget or Upgrade swap from the same row when one exists; specification = the CSV item plus its qty driver. Off-list items get a row with notes: "off-list, cheat sheet pricing". For non-Knox markets, leave material_specs empty unless the user supplies their own SKU pricing.

Generate Reports

Install dependencies if needed:

pip install openpyxl --break-system-packages

1. Save the JSON data to a temp file
2. Run scripts/generate_rehab_excel.py <output_path> <json_path> to create the Excel workbook

Step 7: Present to User

Present the key findings in conversation BEFORE linking the files. Structure:

Summary Format

## Property: [Address]

### Condition Highlights
[3-5 most significant findings from photo/inspection analysis]

### Cost Estimates (Local pricing for [City, State])

| | Full Rehab | Wholetail |
|---|---|---|
| **Estimate** | $XX,XXX | $XX,XXX |
| **$/SF** | $XX.XX | $XX.XX |

### Full Deal Analysis (if ARV/purchase price available)
| Cost Category | Full Rehab | Wholetail |
|---------------|-----------|-----------|
| ARV (Sale Price) | $XXX,XXX | $XXX,XXX |
| Purchase Price | $XXX,XXX | $XXX,XXX |
| Rehab Cost | $XX,XXX | $XX,XXX |
| Financing Costs | $X,XXX | $X,XXX |
| Holding Costs (X mo) | $X,XXX | $X,XXX |
| Buying Costs | $X,XXX | $X,XXX |
| Selling Costs | $XX,XXX | $XX,XXX |
| **Total All-In Cost** | **$XXX,XXX** | **$XXX,XXX** |
| **Net Profit** | **$XX,XXX** | **$XX,XXX** |
| **ROI** | **XX%** | **XX%** |
| MAO (75% Rule) | $XXX,XXX | — |

### Top 5 Budget Line Items
[Largest cost items to set expectations]

### Recommendations
[2-3 strategic observations]

Then link the deliverable:

• Excel workbook (itemized breakdown with all sheets)

Contingency Guidelines

Contingency is applied to the TOTAL rehab cost (not individual line items). It covers items you cannot see in photos: behind walls, under floors, in the crawlspace, and the inevitable scope creep that happens on every project.

Contingency Application Rules

• 15% contingency (standard): Well-inspected properties, standard investor flips with good photo coverage. This is the minimum for most deals.
• 20% contingency (recommended default): Older homes (pre-1970), limited inspection access, any property where you have not personally walked it. Real investors use 20% as their standard.
• 25% contingency (high risk): Unknown conditions, sight-unseen purchases, heavily distressed properties, properties with known structural issues or environmental hazards.
• Apply to TOTAL rehab cost -- never to individual line items. Example: $46,720 subtotal x 20% = $9,344 contingency = $56,064 grand total.
• Only use 10% when the property is newer (post-1990) AND you have comprehensive photos showing good overall condition AND you (or the investor) have walked the property.

Older properties (pre-1970) often have hidden costs: asbestos abatement, lead paint, knob-and-tube wiring, galvanized plumbing, cast iron drain lines. Flag these in the estimate when year_built suggests risk.

Sanity Check — $/SF Benchmarks

After building the estimate (before contingency), divide the subtotal by GLA to get $/SF. Compare against these benchmarks:

If a standard flip estimate exceeds $40/SF, you are almost certainly over-scoping. This is the 60-65% Rule -- go back and ask: "Would a real contractor actually replace this, or would they leave it alone?" Common culprits: including roof speculatively, over-counting windows, full demo when only targeted demo is needed, per-room outlet itemization instead of lump sum.

Rehab Duration Benchmarks

Use these to estimate project timelines and calculate hold time for deal analysis:

Add 2-4 weeks buffer for permitting delays in markets that require permits for rehab work. These timelines assume a single general contractor managing subs -- self-managed rehabs by inexperienced investors typically take 50-100% longer.

Local Pricing Methodology

The Repair Cheat Sheet prices above are calibrated for Tier 2 markets (Midwest -- Dayton, OH baseline). For other markets, apply a regional multiplier.

Key Regional Multipliers

How to apply: Multiply cheat sheet prices by the regional multiplier. Example: LVP at $3.50/SF x 0.88 (Knoxville) = $3.08/SF.

If the user provides their market and it is not listed: Ask if they have local contractor pricing to calibrate against. If not, use the national baseline (1.00) and note that estimates should be verified against 3-5 local contractor SOWs.

Calibration method: Compare the skill's estimate against the first 3-5 actual contractor SOWs received. If estimates consistently run high or low by more than 10%, adjust the regional multiplier accordingly. The multiplier is a living number that improves with real data.

See references/local-pricing-guide.md for the full market multiplier table, BLS labor rate lookup methodology, and climate/regulatory cost factors.

Key Principles

1. Scope like a contractor, not an inspector — Real contractors only replace what's broken, damaged, or too dated to sell. They do NOT replace every outlet, every door, all trim, or every window just because they're in the house. The #1 cause of over-estimation is over-scoping. When in doubt, leave it out.

2. The 60-65% Sanity Check — After building your estimate, divide total by GLA. If $/SF exceeds $40 on a standard flip (no systems work), you are over-scoping. Go back and cut items a real contractor would skip.

3. Local pricing matters — A $30/hr labor market (Dayton, OH) produces fundamentally different budgets than a $65/hr market (San Francisco). Always localize. The calibration file provides Tier 2 (Midwest) benchmarks.

4. Mirror the comps — The rehab finish level should match what sold in Bucket B. Over-improving beyond what comps support is the #1 rookie mistake.

5. Wholetail is a real strategy — Not every deal needs a full rehab. Sometimes $8K in paint, flooring, and cleaning nets more ROI than a $45K full gut. Present both options so the investor can choose.

6. Photos > assumptions — When photos are available, use what you can actually see. When they're not, disclose that estimates are based on age/condition assumptions and widen the contingency.

7. Permits, staging, and photos are NOT contractor costs — Never include these in the renovation estimate. They are separate budget line items in the deal analysis, not the SOW.

8. Paint is one line item — Real contractors quote paint as a single all-in $/SF rate. Never split into walls + trim + ceilings as 3 separate line items.

9. One estimate, not three — Produce a single cost per line item targeting the low-to-mid range. Real contractors give you one number, not a range. Three-tier estimates create analysis paralysis and the mid tends to over-estimate. If there's genuine uncertainty on a specific item (e.g., roof condition unknown), note the uncertainty in the item's notes field and widen the contingency percentage instead.

10. 75% Rule for MAO — Max Allowable Offer = ARV × 75% − Rehab Cost. This replaces the traditional 70% rule to reflect realistic deal economics — the 70% rule is overly conservative and causes investors to pass on viable deals.

11. Rehab cost is NOT profit — A deal's true profit = ARV minus ALL costs: purchase, rehab, financing, holding, buying transaction costs, and selling transaction costs. Financing + holding + transaction costs typically add $25K-40K+ on top of the rehab budget. Showing profit as just "ARV minus purchase minus rehab" gives investors a dangerously inflated view that leads to bad buying decisions. Always calculate and present the full deal cost picture.

12. Conservative means overestimate, not underestimate — Every price in this skill is deliberately set at or above the likely real cost. It is better to budget $4,000 for a bathroom and spend $3,500 than to budget $3,000 and get a $4,000 bill. The contingency percentage exists to cover surprises, but the base prices themselves should not require the contingency to be accurate. This protects investor margins and prevents bad buying decisions.

13. 4-tier system matches source code — The finish tiers are: Tier 1 (Minimum Viable), Tier 2 (Builder Grade), Tier 3 (Mid-Range Upgrade), Tier 4 (Premium/Custom). These names match the TIER_NAMES dictionary in rehab_estimator.py. When referencing tiers in output, use these exact names.

14. The locked master material list is the material source on Knox deals. The team locked in data/master_material_list_37914.csv (88 Home Depot SKUs + 12 allowances, exact prices at delivery zip 37914). For Knoxville / Knox / Blount properties, the vast majority of material selections and material prices come from that list; go off-list only for an outstanding random issue the list does not carry, and flag it. Never apply the regional multiplier to locked material prices (they are already local); the multiplier stays on labor. Cheat-sheet lump sums still govern the BUDGET numbers and every non-Knox market; the locked list is what makes the material spec and shopping list exact.

After the estimate: who does the work

A scope of work is only as good as the crew that executes it. When the user needs contractors or subs for the market (new market, a flaked GC, a specialty trade like septic or a new water line), route to the vendor-directory-builder skill: it mines the local investor community, verifies every provider against public records, and ships a vetted directory. The contractor-call-sheet skill then turns that directory into a call sheet and first-contact drafts.