financial-plan-builder

name	financial-plan-builder
description	Build a 12-month cash projection from a YAML config (starting balance, revenue assumptions, fixed costs, variable costs, hiring plan). Outputs a monthly cash table, runway in months, break-even month, and a simple sensitivity analysis. Use when planning runway, deciding to raise / not raise, modeling a hire, or producing a board-update spreadsheet. NOT financial advice. Numbers, not opinions.

Financial-plan-builder

A spreadsheet that lives in version control. Reads a YAML config, emits a monthly cash projection with the numbers you actually need to make funding-vs-runway decisions.

Disclaimer

This skill produces numbers from a YAML config. It does not:

• Advise on whether to raise capital.
• Predict revenue.
• Account for taxes (the skill emits a [TAX] flag wherever tax treatment matters; the operator/accountant fills in).
• Model complex equity, debt, or warrant structures.
• Replace an accountant or financial controller for material decisions (audited reports, tax filings, due diligence).

If the YAML inputs are wrong, the output is wrong. The skill's value is consistency — the same config produces the same projection every time, so the operator can iterate on the inputs (what if I hire one engineer in month 4 instead of month 2?) without rebuilding the spreadsheet.

Input — finances/plan.yml

The skill reads finances/plan.yml at the repo root (or a path the operator passes). Example:

plan:
  name: 2026 H1 projection
  starting_balance_usd: 180000        # cash in bank on day 1
  start_date: 2026-01-01
  horizon_months: 12

revenue:
  # MRR streams. The skill compounds growth at the configured rate.
  - name: self-serve-saas
    starting_mrr_usd: 4500            # MRR at start_date
    monthly_growth_rate: 0.08         # 8% MoM
    cap_mrr_usd: 30000                # plateau cap

  - name: enterprise-contracts
    starting_mrr_usd: 0
    contracts:                        # discrete deals, not MoM growth
      - start_month: 3
        monthly_value_usd: 8000
        term_months: 12
      - start_month: 7
        monthly_value_usd: 12000
        term_months: 6

  - name: consulting
    starting_mrr_usd: 0
    one_off_payments:                 # non-recurring lumps
      - month: 2
        usd: 15000
      - month: 5
        usd: 10000

fixed_costs:
  # Costs that don't scale with revenue — paid every month.
  rent: 0                              # remote-first
  software_subscriptions: 850          # Notion, Linear, Vercel, etc.
  hosting: 320                         # AWS / Hetzner / Cloudflare
  legal_accounting_fixed: 400          # retainer or monthly accountant
  insurance: 200
  other: 100

variable_costs:
  # Costs that scale with usage or revenue. Skill computes from
  # configured percentages.
  payment_processing_pct: 0.029        # Stripe-shape
  llm_api_pct_of_mrr: 0.04             # AI API as % of MRR (operator estimates)
  cogs_pct_of_revenue: 0.10            # other variable costs

hiring_plan:
  # Each entry adds to monthly burn at the start_month.
  - role: founder
    start_month: 0
    monthly_total_usd: 6000            # salary + benefits + employer-side tax
    notes: operator self-pay

  - role: engineer
    start_month: 4
    monthly_total_usd: 11000

  - role: marketing
    start_month: 8
    monthly_total_usd: 6500
    notes: contractor, half-time

one_off_costs:
  # Lumpy expenses outside the recurring categories.
  - description: trademark filing
    month: 3
    usd: 2200

  - description: conference + travel
    month: 6
    usd: 4800

sensitivity:
  # The skill runs three additional scenarios besides the base case.
  - name: bull
    revenue_multiplier: 1.30           # all revenue lines * 1.30
    cost_multiplier: 1.00

  - name: bear
    revenue_multiplier: 0.70
    cost_multiplier: 1.15               # costs run hot

  - name: bear-and-hire-paused
    revenue_multiplier: 0.70
    cost_multiplier: 1.15
    skip_hires: [engineer, marketing]   # postpone these

Output

Writes to finances/<plan-name>/:

finances/<plan-name>/
├── plan.yml                       (the input — kept for traceability)
├── projection.md                  (human-readable monthly table)
├── projection.csv                 (importable into a spreadsheet)
└── summary.md                     (the 1-page board update)

projection.md shape

# Projection: 2026 H1

| Month | Date | Revenue | Costs | Net | Cash balance | Runway |
|---|---|---|---|---|---|---|
| 0 | 2026-01 | $4,500 | $14,000 | -$9,500 | $170,500 | 18 mo |
| 1 | 2026-02 | $4,860 | $14,200 | -$9,340 | $161,160 | 17 mo |
| 2 | 2026-03 | $20,249 | $14,500 | $5,749 | $166,909 | n/a |
...
| 11 | 2026-12 | $42,650 | $28,800 | $13,850 | $228,150 | n/a |

**Starting balance**: $180,000
**Ending balance (month 11)**: $228,150
**Cumulative net**: +$48,150
**Break-even month**: 2 (2026-03)
**Lowest cash point**: month 1 ($161,160)
**Runway from month 0 (cost-only)**: 19 months
**Runway from lowest point (cost-only)**: 17 months

summary.md (board-update grade)

# Q1 2026 — financial summary

**Starting cash**: $180,000
**Plan horizon**: 12 months

## Headline numbers

- Ending cash (month 11): **$228,150**
- Break-even: **month 2** (2026-03)
- Lowest cash point: **$161,160** (month 1)
- Monthly burn at month 0: **$9,500**
- Monthly contribution at month 11: **+$13,850**

## Sensitivity

| Scenario | Ending cash | Lowest point | Break-even |
|---|---|---|---|
| Base | $228,150 | $161,160 | month 2 |
| Bull (rev × 1.30) | $341,000 | $164,200 | month 1 |
| Bear (rev × 0.70, costs × 1.15) | $32,400 | $32,400 | month 11 |
| Bear + hire pause | $98,600 | $73,200 | month 10 |

## Key risks (operator review required)

- Bear scenario ends with $32k cash — below 3-month buffer.
- Engineer hire in month 4 is the single largest cost commit.
- Self-serve growth assumes 8% MoM; historical was 5%. [REVIEW]

## What this projection does NOT include

- Taxes (federal, state, sales). See `[TAX]` markers in `projection.md`.
- Equity dilution from a future raise.
- Variability in payment processing rate.
- Bad-debt allowance.

What the skill actually computes

For each month m from 0 to horizon_months - 1:

1. Revenue:

   • For each MRR line: starting MRR × (1 + growth_rate) ^ m, capped at cap_mrr_usd.
   • Plus contracts active in month m (start_month ≤ m < start_month + term).
   • Plus one-off payments scheduled for month m.

2. Fixed costs: sum of fixed_costs.* (constant per month).

3. Variable costs: revenue × (payment_processing_pct + llm_api_pct_of_mrr × MRR/revenue + cogs_pct_of_revenue). The LLM term is treated as a function of MRR specifically; one-off payments don't trigger LLM cost.

4. Hiring costs: sum of hiring_plan[i].monthly_total_usd for each role with start_month ≤ m.

5. One-off costs: cost from one_off_costs[i] if i.month == m, else 0.

6. Net: revenue - (fixed + variable + hiring + one_off).

7. Cash balance: cumulative sum of net + starting_balance.

8. Runway: if monthly net < 0, runway = current_balance / |net|. Else "n/a".

For sensitivity scenarios: multiply revenue / cost streams by the configured multipliers, then re-run the computation. skip_hires removes those roles entirely from the plan.

What this skill is NOT

• A forecasting tool. The skill projects the YAML inputs; it does not predict revenue from anything other than what the operator configures. If the operator's assumptions are wrong, the projection is wrong.
• A bookkeeping tool. The skill does not reconcile against actual bank statements. Bookkeeping happens in QuickBooks / Xero / a spreadsheet; this skill answers "what does the next 12 months look like if we keep going".
• A tax planning tool. [TAX] markers flag where the operator's accountant needs to weigh in.
• An ERP / billing system. The skill does not invoice, dunning, reconcile, or collect.

Operator workflow

1. First run: scaffold finances/plan.yml from the template above. Adjust starting balance + revenue lines + hiring plan to match reality.
2. Re-run any time an assumption changes (close a deal, hire, raise prices). Commit the new plan.yml so the projection history is version-controlled.
3. Monthly cadence: open summary.md, compare projected vs. actual, adjust assumptions for any line that drifted by > 10%.
4. Board update: paste summary.md content into the board email / deck. Highlight the sensitivity row that aligns with this month's experience.

Anti-patterns

1. Treating the bull case as the plan. Plan against base or bear; if bull happens, you have more options. Planning against bull and getting base = layoffs.
2. Ignoring [TAX] markers. US LLCs pass through to operator income; C-corps pay corporate; international LLCs vary widely. Talk to an accountant.
3. Hiring on the assumption next month's revenue clears the payroll commitment. Hire when you have 6+ months of runway AFTER the salary commitment is added.
4. Updating the YAML retroactively to match what actually happened. That makes the historical projections look prescient when they weren't. Keep an "actuals" alongside the projection for honest comparison.

Hard rules

• Always emit [TAX] markers where taxes apply but aren't modeled.
• Always show the lowest cash point, not just the ending balance. A plan that ends at $500k but dips to $5k in month 7 is not a safe plan.
• Never assume revenue growth above 15% MoM without an explicit [REVIEW] marker. Sustained 15%+ MoM is exceptional; the skill warns when configured.
• The output is a Markdown table + CSV, not a chart. Charts belong in a separate visualization layer. Numbers first.

Integration with other Walter-OS skills

• hiring-toolkit — when planning a new hire, this skill models the cost impact; that skill produces the JD / interview rubric / offer letter for the same role.
• pricing-experiment — when changing prices, this skill models the revenue impact under different MRR growth assumptions.
• vc-evaluator — when prepping for a raise, this skill produces the projection that goes into the deck; that skill helps anticipate investor pushback on the numbers.
• saas-metrics-dashboard — for the metrics layer (MRR, ARR, churn, LTV/CAC); this skill projects forward, that skill measures backward.

References

• Walter-OS execution plan Phase 4.1.4.
• skills/saas-metrics-dashboard/SKILL.md — backward-looking metrics; this skill is the forward-looking complement.
• skills/hiring-toolkit/SKILL.md — JD/rubric/offer-letter for the roles in the hiring plan.