Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Build a comparative market analysis (CMA) to price a property. Use when asked to do a CMA, a comparative market analysis, price a home, or estimate a property's value from comparables. Produces a structured CMA — the subject property, selected comparables with adjustments, an estimated value range, market context, and a pricing recommendation with rationale — for a real-estate professional to review. Not a formal appraisal.
.claude/skills/mohitagw15856-comparative-market-analysis/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -4% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -15% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 123% | 0% |
| case-15 | ✗→✓ | ▲ Improved | 22% | 0% |
| case-04 | ✓→✓ | = Same ✓ | 113% | 0% |
A CMA prices a home the way the market actually values it: against recent, similar, nearby sales — adjusted for the differences. This skill structures that analysis so the number is defensible: the comparables chosen and why, the adjustments made, the resulting range, and a pricing recommendation tied to the seller's goal.
> Note: this is a pricing-analysis aid for a real-estate professional, not a formal appraisal or > financial/legal advice. It works from the comparables and figures you provide; valuation depends on local > market data and professional judgement. Never invent comp sales or prices — use the data given or mark it to source.
Given a subject property and a few comps, build the CMA anyway — structure the analysis, apply reasoned adjustments, and give a range, marking any figure to source (confirm with MLS/records). Where comps are missing, explain what to pull rather than inventing sales. Never fabricate comparable prices.
Ask for these only if they aren't already provided (else mark to source):
1. Subject property — the key attributes summarised.
2. Comparables — a table of the comps used, with adjustments toward the subject:
| Comp | Sold price | Date | Beds/Baths | Size | Key differences | Adjustment | Adjusted price | |---|---|---|---|---|---|---|---|
Explain the adjustment logic (e.g. +/- for size, condition, extra bath, garage, view) — directionally and why.
3. Market context — the trend, inventory, and absorption, and what it means for pricing now.
4. Estimated value range — a supported range from the adjusted comps (not a single false-precision number), with the most-likely figure.
5. Pricing recommendation — a list price tied to the goal (e.g. price at market for speed, slightly under for multiple offers, at the top of range to test) — with the trade-off of each.
6. Caveats — data to confirm, and a note that a formal appraisal/agent review is needed.
Real-estate valuation practice — comparable-sales analysis with feature adjustments, market-context weighting, and goal-aligned pricing (CMA, not a formal appraisal).
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 18,554 | 13,173 | -29% | 1 | 1 | 0% | 3,763 | 3,599 | -4% | 0 | 0 | — |
case-02 | fail→fail | 15,101 | 15,069 | -0% | 1 | 1 | 0% | 3,192 | 4,269 | +34% | 0 | 0 | — |
case-03 | fail→fail | 16,420 | 12,606 | -23% | 1 | 1 | 0% | 3,588 | 3,618 | +1% | 0 | 0 | — |
case-04 | pass→pass | 6,024 | 8,322 | +38% | 1 | 1 | 0% | 1,021 | 2,170 | +113% | 0 | 0 | — |
case-05 | fail→pass | 14,020 | 7,499 | -47% | 1 | 1 | 0% | 2,526 | 2,152 | -15% | 0 | 0 | — |
case-06 | fail→fail | 17,704 | 16,591 | -6% | 1 | 1 | 0% | 3,508 | 4,158 | +19% | 0 | 0 | — |
case-07 | pass→pass | 6,587 | 12,128 | +84% | 1 | 1 | 0% | 1,321 | 2,907 | +120% | 0 | 0 | — |
case-08 | fail→pass | 6,569 | 9,449 | +44% | 1 | 1 | 0% | 1,216 | 2,714 | +123% | 0 | 0 | — |
case-09 | pass→pass | 10,767 | 10,205 | -5% | 1 | 1 | 0% | 2,278 | 2,861 | +26% | 0 | 0 | — |
case-10 | pass→pass | 8,070 | 10,013 | +24% | 1 | 1 | 0% | 1,432 | 2,813 | +96% | 0 | 0 | — |
case-11 | pass→pass | 10,337 | 9,579 | -7% | 1 | 1 | 0% | 1,695 | 2,606 | +54% | 0 | 0 | — |
case-12 | pass→pass | 12,776 | 10,319 | -19% | 1 | 1 | 0% | 2,264 | 2,779 | +23% | 0 | 0 | — |
case-13 | pass→pass | 10,622 | 9,298 | -12% | 1 | 1 | 0% | 2,040 | 2,719 | +33% | 0 | 0 | — |
case-14 | pass→pass | 11,500 | 10,902 | -5% | 1 | 1 | 0% | 1,966 | 2,745 | +40% | 0 | 0 | — |
case-15 | fail→pass | 7,201 | 4,952 | -31% | 1 | 1 | 0% | 1,339 | 1,627 | +22% | 0 | 0 | — |
case-16 | pass→pass | 13,549 | 12,352 | -9% | 1 | 1 | 0% | 3,133 | 3,574 | +14% | 0 | 0 | — |
case-17 | pass→pass | 10,402 | 10,655 | +2% | 1 | 1 | 0% | 1,906 | 2,956 | +55% | 0 | 0 | — |
case-18 | pass→pass | 10,419 | 8,892 | -15% | 1 | 1 | 0% | 2,072 | 2,548 | +23% | 0 | 0 | — |
case-19 | pass→pass | 12,738 | 13,197 | +4% | 1 | 1 | 0% | 2,278 | 3,193 | +40% | 0 | 0 | — |
case-20 | fail→fail | 13,893 | 12,219 | -12% | 1 | 1 | 0% | 2,543 | 3,112 | +22% | 0 | 0 | — |
case-21 | pass→pass | 4,821 | 9,531 | +98% | 1 | 1 | 0% | 1,174 | 2,632 | +124% | 0 | 0 | — |
case-22 | pass→pass | 8,847 | 10,374 | +17% | 1 | 1 | 0% | 1,503 | 2,617 | +74% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +18 percentage points is the difference between those two pass rates over the 22 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.