Financial Model

List all inputs needed:

• Existing budget files (check library/docs/research/ for prior finance reports)
• Known fixed costs (subscriptions, tooling, headcount)
• Variable costs (usage-based APIs, campaign spend)
• Revenue or income data if applicable

Never invent numbers. If a data point is missing, flag it as an assumption with a range estimate.

Step 3 — Structure the Model

Use this standard layout for any model:

INPUTS (facts and confirmed data)
├── Revenue: ...
├── Fixed costs: ...
└── Variable costs: ...

ASSUMPTIONS (estimates with rationale)
├── Growth rate: X% — based on [source/rationale]
└── Churn rate: Y% — based on [source/rationale]

CALCULATIONS
├── Gross margin: Revenue − COGS
├── Operating expenses: Fixed + Variable
└── Net: Gross margin − OpEx

OUTPUTS
├── Monthly burn rate
├── Runway (months)
└── ROI or payback period (if applicable)

Step 4 — Run Scenarios

For any recommendation, produce 3 scenarios:

• Conservative: Pessimistic assumptions (worst-case inputs)
• Base: Most likely assumptions
• Optimistic: Best-case inputs

Label each scenario clearly. Never present only the base case.

Step 5 — Validate

Before finalizing:

• Do totals add up? Verify arithmetic.
• Are all assumptions explicitly stated and defensible?
• Are currency, units, and time periods consistent throughout?
• Flag any input that has >20% uncertainty as HIGH RISK.

Step 6 — Approval Gate

Any recommendation involving fund movement or budget reallocation:

1. Draft recommendation with rationale
2. Mark task as human-review — do NOT approve your own financial recommendations
3. Include: amount, purpose, expected ROI, risk level

Step 7 — Save & Report

• Save to library/docs/research/YYYY-MM-DD_finance_<description>.md
• Strategy docs to library/docs/stratagies/YYYY-MM-DD_strategy_<description>.md
• Include data range, sources, and scenario summary at top of file

Output Format

## Financial Analysis: [Title]
Date: YYYY-MM-DD
Period: [Q1 2025 / FY 2025 / etc.]
Question: [What this answers]

### Summary
[3-sentence executive summary with the key number and recommendation]

### Inputs
| Item | Value | Source |
|------|-------|--------|
| ... | ... | ... |

### Assumptions
| Assumption | Value | Rationale |
|------------|-------|-----------|
| ... | ... | ... |

### Scenarios
| Scenario | Monthly Burn | Runway | Net |
|----------|-------------|--------|-----|
| Conservative | ... | ... | ... |
| Base | ... | ... | ... |
| Optimistic | ... | ... | ... |

### Recommendation
[Approve / Deny / Conditional — with clear reasoning]

### Risk Flags
- [Any HIGH RISK inputs or assumptions]

Examples

Good task for this skill: "Analyze Q1 API costs and model what happens if usage doubles in Q2."

Good task for this skill: "Is $500/month on the new analytics tool justified by the time it saves?"

Escalation trigger: Any recommendation over $1,000 or touching treasury funds → always route to human-review. Finance Manager never self-approves.