Install any skill in seconds. Free to start, no credit card required.
Get Started Free →When you want to model personal financial scenarios — house purchase + rental income (ADU, bedroom rentals, house-hacking), renovation budgets, monthly cash flow forecasts, big-purchase decisions, savings/investment what-ifs. For personal life: a household (you + partner), household budgets, real-estate decisions. v0.1 ships with the house scenario template (purchase + rental scenarios) as the first use case. Architected so other personal-finance scenarios (refi, car, education, retirement, side
.claude/skills/coreyhaines31-personal-cfo/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-08 | ✗→✓ | ▲ Improved | 181% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 89% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 103% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 37% | 0% |
| case-13 | ✗→✓ | ▲ Improved | 150% | 0% |
This skill is for a household (you + partner): housing decisions, monthly cash flow, big purchases, scenario what-ifs. Pair it with a business-books CFO skill if you also run one for a company — the DNA is the same, the domain is different.
| Invocation | Type | What it models | |---|---|---| | /personal-cfo house (default for now) | house | Buy a house + rental income scenarios (ADU / bedrooms / combinations) + renovations + monthly cash flow | | /personal-cfo refi | refi | Refinancing scenarios | | /personal-cfo cash-flow | cash-flow | Generic monthly cash flow (income, expenses, savings rate) | | /personal-cfo big-purchase | big-purchase | Should we buy X (car, education, equipment) — cash vs finance vs lease comparisons | | /personal-cfo side-income | side-income | Adding a side income stream — break-even, tax implications | | /personal-cfo savings-what-if | savings-what-if | What if we put $X into Y for N years — index funds, real estate, etc. |
For v0.1, only the house template is fleshed out. Others are sketched in references/templates/ as stubs.
Ask in a structured batch. For house mode, see references/templates/house.md for the full input questionnaire. Core fields:
Property
Renovations / repairs
Rental scenarios (for the house mode)
Personal
If anything's fuzzy, ask. Don't proceed with placeholder numbers — financial models with wrong inputs are dangerous.
For the house mode, default scenarios to compare:
| Scenario | Description | |---|---| | Baseline | Own + live in, no rental income | | A: ADU only | Rent the ADU, live in main house | | B: 1 bedroom only | House-hack 1 bedroom, keep ADU empty (or for guests) | | C: 2 bedrooms only | House-hack 2 bedrooms, ADU empty | | D: ADU + 1 bedroom | Both ADU and 1 bedroom rented | | E: ADU + 2 bedrooms | Maximum rental | | F: ADU + bedrooms (year 2+) | Year 1: just ADU while renovating. Year 2+: add bedrooms after partner's comfort |
Let the user customize the scenario list — every house is different.
For each scenario, calculate:
Monthly
references/calculators.md for the formulaAnnual
Long-term
Output to ~/Documents/personal-cfo/<scenario-slug>-<YYYY-MM-DD>/:
inputs.yaml — all the input numbers (so the analysis is rerunnable + auditable)report.md — the full analysis with assumptions called outcomparison.md — side-by-side scenario table (the headline output)comparison.md templatemarkdown# <Property nickname> — scenario comparison **Date:** YYYY-MM-DD **Property:** <address or nickname> **Time horizon:** 10 years **Key assumptions:** 3%/yr appreciation, 7% S&P alt return, 6% vacancy | Scenario | Net monthly cost | Annual cost | 10-yr total | Equity @ 10yr | Notes | |---|---:|---:|---:|---:|---| | Baseline | $X,XXX | $XX,XXX | $XXX,XXX | $XXX,XXX | No rental income | | A: ADU | $X,XXX | $XX,XXX | $XXX,XXX | $XXX,XXX | $X,XXX/mo ADU rent | | B: 1 BR | … | … | … | … | … | | C: 2 BR | … | … | … | … | … | | D: ADU + 1 BR | … | … | … | … | … | | E: ADU + 2 BR | … | … | … | … | … | ## Headline takeaway <1–2 sentences: which scenario wins on cash flow, which wins on lifestyle, where the inflection points are> ## Sensitivity analysis - If rates rise 1%: <impact> - If ADU rent comes in 20% lower: <impact> - If renovations overrun 30%: <impact> - If vacancy doubles to 12%: <impact> ## Open questions - <what would change the recommendation> - <what to research next — composes with /deep-research>
Save the scenario folder so future revisits work. The archive index lives in ${MAKERSKILLS_CONFIG:-$HOME/.config/makerskills}/personal-cfo/archive/ (create the directory if missing). Never write archives inside the skill's own folder — skill installs and upgrades re-sync from source and wipe anything saved there. Migration: if this skill's folder contains an old references/scenarios-archive/ with user entries, move those files into the archive directory first.
Append a one-liner to <archive dir>/INDEX.md (create if missing):
markdown- 2026-06-26 — [<property nickname>](~/Documents/personal-cfo/<slug>-2026-06-26/) — <type> — <headline takeaway>
This compounds — when you're considering Property #2 a year later, you can reuse the model and grep past analyses for patterns.
/deep-research on rental comps for the area to firm up the ADU and bedroom rent assumptions?"/decide once the scenarios land?"second-brain raw/ as note-house-<slug>.md?"decide — once scenarios are modeled, the buy / don't / negotiate decision goes through /decide. Cash flow analysis is input, not the decision itself.deep-research — for inputs that need market data: rental comps, mortgage rate trends, neighborhood trajectory, school ratings, comparable sales.business-brainstorm — if the scenario tips into "this is a business" (e.g., short-term rental, multi-unit purchase, scaling to N properties), route there to pressure-test the business angle.second-brain — capture analyses to raw/note-house-<slug>.md so they live in the personal wiki and can be referenced later./decide makes the call.refi.md — refinance scenarioscash-flow.md — generic monthly budgetbig-purchase.md — cash vs finance vs lease for big-ticket itemsside-income.md — adding a side income streamsavings-what-if.md — investment what-ifsv0.1 ships only house.md fleshed out. Add others when the use case is real.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 26,873 | 25,217 | -6% | 1 | 1 | 0% | 5,003 | 7,269 | +45% | 0 | 0 | — |
case-02 | fail→fail | 26,301 | 6,529 | -75% | 1 | 1 | 0% | 5,046 | 3,505 | -31% | 0 | 0 | — |
case-03 | fail→fail | 39,949 | 7,996 | -80% | 1 | 1 | 0% | 6,256 | 3,624 | -42% | 0 | 0 | — |
case-04 | fail→fail | 12,977 | 19,266 | +48% | 1 | 1 | 0% | 2,398 | 6,047 | +152% | 0 | 0 | — |
case-05 | pass→pass | 14,179 | 8,893 | -37% | 1 | 1 | 0% | 2,079 | 3,703 | +78% | 0 | 0 | — |
case-06 | pass→pass | 12,522 | 6,852 | -45% | 1 | 1 | 0% | 1,878 | 3,343 | +78% | 0 | 0 | — |
case-07 | pass→pass | 13,943 | 10,063 | -28% | 1 | 1 | 0% | 2,166 | 3,791 | +75% | 0 | 0 | — |
case-08 | fail→pass | 7,573 | 4,253 | -44% | 1 | 1 | 0% | 1,122 | 3,151 | +181% | 0 | 0 | — |
case-09 | fail→pass | 10,530 | 3,811 | -64% | 1 | 1 | 0% | 1,567 | 2,963 | +89% | 0 | 0 | — |
case-10 | fail→pass | 10,061 | 3,062 | -70% | 1 | 1 | 0% | 1,412 | 2,864 | +103% | 0 | 0 | — |
case-11 | fail→pass | 17,916 | 10,274 | -43% | 1 | 1 | 0% | 2,767 | 3,798 | +37% | 0 | 0 | — |
case-12 | pass→pass | 13,315 | 4,941 | -63% | 1 | 1 | 0% | 2,116 | 3,081 | +46% | 0 | 0 | — |
case-13 | fail→pass | 8,513 | 5,303 | -38% | 1 | 1 | 0% | 1,243 | 3,107 | +150% | 0 | 0 | — |
case-14 | fail→pass | 16,213 | 10,470 | -35% | 1 | 1 | 0% | 2,408 | 3,927 | +63% | 0 | 0 | — |
case-15 | fail→pass | 12,241 | 6,579 | -46% | 1 | 1 | 0% | 1,902 | 3,570 | +88% | 0 | 0 | — |
case-16 | fail→pass | 13,297 | 6,110 | -54% | 1 | 1 | 0% | 2,146 | 3,493 | +63% | 0 | 0 | — |
case-17 | pass→pass | 12,020 | 8,024 | -33% | 1 | 1 | 0% | 2,110 | 3,617 | +71% | 0 | 0 | — |
case-18 | pass→pass | 8,001 | 5,353 | -33% | 1 | 1 | 0% | 1,429 | 3,201 | +124% | 0 | 0 | — |
case-19 | pass→pass | 10,973 | 5,499 | -50% | 1 | 1 | 0% | 1,740 | 3,145 | +81% | 0 | 0 | — |
case-20 | fail→fail | 23,770 | 15,277 | -36% | 1 | 1 | 0% | 4,555 | 4,811 | +6% | 0 | 0 | — |
case-21 | pass→pass | 8,619 | 7,612 | -12% | 1 | 1 | 0% | 1,268 | 3,560 | +181% | 0 | 0 | — |
case-22 | pass→pass | 9,271 | 5,831 | -37% | 1 | 1 | 0% | 1,324 | 3,210 | +142% | 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 +36 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.