Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Verify Fondo setup by checking financial data sync, reviewing categorized transactions, and confirming R&D tax credit eligibility. Trigger: "fondo first sync", "fondo verify", "fondo hello world", "fondo test".
.claude/skills/jeremylongshore-fondo-hello-world/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -3% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -37% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -7% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -41% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -52% | 0% |
Verify your Fondo setup is working: check that bank transactions are syncing, payroll data is flowing, and your company qualifies for R&D tax credits.
Navigate to Fondo Dashboard > Transactions:
Fondo automatically categorizes transactions:
| Category | Examples | Tax Impact | |----------|----------|------------| | Payroll | Gusto payments, contractor 1099s | Deductible, R&D qualified | | Software | AWS, GitHub, Figma | Deductible, R&D qualified | | Office | WeWork, office supplies | Deductible | | Travel | Flights, hotels, meals | Partially deductible | | Revenue | Stripe payouts, customer payments | Taxable income | | Transfers | Between own accounts | Not taxable |
Review and correct any miscategorized transactions in Dashboard > Transactions.
Fondo Dashboard > Tax Credits > R&D Assessment
Eligible if ALL apply:
✓ US-based employees (W-2, not just contractors)
✓ Developing new/improved products, processes, or software
✓ Technical uncertainty exists in the development
✓ Systematic experimentation/iteration to resolve uncertainty
Average startup credit: $21,000/year
Maximum (payroll tax offset): $500,000/yearDashboard > Payroll Integration
→ Confirm employee count matches your payroll provider
→ Verify salary data for R&D credit calculations
→ Check contractor payments are categorized separatelyAfter 48 hours of setup:
| Issue | Solution | |-------|----------| | No transactions appearing | Check Plaid connection status, allow 24-48h | | Wrong categorization | Manually recategorize, Fondo learns patterns | | R&D credit shows $0 | Verify W-2 employees doing qualifying work | | Payroll data missing | Re-authorize payroll provider OAuth |
For development workflow details, see fondo-local-dev-loop.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 19,080 | 13,633 | -29% | 1 | 1 | 0% | 2,292 | 2,232 | -3% | 0 | 0 | — |
case-02 | fail→pass | 19,075 | 4,363 | -77% | 1 | 1 | 0% | 2,406 | 1,508 | -37% | 0 | 0 | — |
case-03 | fail→pass | 14,401 | 12,955 | -10% | 1 | 1 | 0% | 2,256 | 2,094 | -7% | 0 | 0 | — |
case-04 | pass→pass | 18,751 | 8,579 | -54% | 1 | 1 | 0% | 2,161 | 1,293 | -40% | 0 | 0 | — |
case-05 | pass→fail | 15,950 | 11,656 | -27% | 1 | 1 | 0% | 1,797 | 1,773 | -1% | 0 | 0 | — |
case-06 | pass→pass | 19,178 | 8,584 | -55% | 1 | 1 | 0% | 2,662 | 2,274 | -15% | 0 | 0 | — |
case-07 | pass→pass | 12,401 | 3,437 | -72% | 1 | 1 | 0% | 2,006 | 1,200 | -40% | 0 | 0 | — |
case-08 | fail→pass | 19,964 | 9,359 | -53% | 1 | 1 | 0% | 2,439 | 1,443 | -41% | 0 | 0 | — |
case-09 | fail→pass | 16,236 | 6,951 | -57% | 1 | 1 | 0% | 1,975 | 944 | -52% | 0 | 0 | — |
case-10 | pass→pass | 14,936 | 3,390 | -77% | 1 | 1 | 0% | 1,547 | 1,165 | -25% | 0 | 0 | — |
case-11 | pass→pass | 12,421 | 7,067 | -43% | 1 | 1 | 0% | 1,211 | 908 | -25% | 0 | 0 | — |
case-12 | fail→pass | 12,140 | 9,523 | -22% | 1 | 1 | 0% | 2,017 | 1,404 | -30% | 0 | 0 | — |
case-13 | fail→pass | 11,833 | 7,048 | -40% | 1 | 1 | 0% | 1,181 | 940 | -20% | 0 | 0 | — |
case-14 | fail→pass | 15,515 | 7,941 | -49% | 1 | 1 | 0% | 1,901 | 1,072 | -44% | 0 | 0 | — |
case-15 | fail→pass | 9,337 | 7,646 | -18% | 1 | 1 | 0% | 1,519 | 1,099 | -28% | 0 | 0 | — |
case-16 | fail→pass | 11,257 | 6,632 | -41% | 1 | 1 | 0% | 1,083 | 883 | -18% | 0 | 0 | — |
case-17 | fail→fail | 8,844 | 8,680 | -2% | 1 | 1 | 0% | 1,402 | 1,178 | -16% | 0 | 0 | — |
case-18 | pass→pass | 15,252 | 5,280 | -65% | 1 | 1 | 0% | 2,610 | 1,574 | -40% | 0 | 0 | — |
case-19 | pass→pass | 23,555 | 15,910 | -32% | 1 | 1 | 0% | 3,829 | 4,105 | +7% | 0 | 0 | — |
case-20 | pass→pass | 17,595 | 15,675 | -11% | 1 | 1 | 0% | 2,402 | 2,547 | +6% | 0 | 0 | — |
case-21 | pass→pass | 18,085 | 21,349 | +18% | 1 | 1 | 0% | 3,362 | 3,563 | +6% | 0 | 0 | — |
case-22 | fail→pass | 15,803 | 6,317 | -60% | 1 | 1 | 0% | 1,826 | 810 | -56% | 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 +45 percentage points is the difference between those two pass rates over the 22 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
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.