| name | financial-tracker-ops |
| description | Maintains ledgers, cashflow projections, debt schedules, budgets, and reconciliation artifacts using double-entry bookkeeping principles. Trigger phrases: "cashflow forecast", "budget vs actual", "debt payoff", "ledger entry", "amortization schedule", "reconciliation", "13-week forecast", "variance analysis", "balance sheet", "income statement", "accounts receivable". Do NOT use for: market sizing or industry research (use market-research), building pivot tables or data transforms on raw spreadsheets (use spreadsheet-analysis), or investment portfolio analysis beyond basic tracking. |
| license | Apache-2.0 |
| compatibility | {"clients":["openai-codex","gemini-cli","opencode","github-copilot"]} |
| metadata | {"owner":"codex","domain":"financial-tracker-ops","maturity":"draft","risk":"low","tags":["finance","bookkeeping","cashflow","budgeting","debt-tracking"]} |
Purpose
Create and maintain financial tracking artifacts — ledgers, cashflow forecasts, debt schedules, budgets, and reconciliation reports — using double-entry bookkeeping principles. Every financial artifact must balance, state its assumptions, and maintain an audit trail.
When to use this skill
- The user asks to record transactions, maintain a ledger, or create journal entries.
- A task requires cashflow projection or a 13-week rolling forecast.
- The user needs debt tracking: amortization schedules, payoff comparisons (avalanche vs. snowball), or loan analysis.
- Budget-vs-actual variance analysis is requested.
- A repo contains financial tracking files (e.g.,
finances/ledger.csv, budget/2024-Q3.md) that need updating or auditing.
- Account reconciliation is needed — matching ledger entries against bank statements or external records.
Do not use this skill when
- The task is market sizing, competitive analysis, or industry research — use
market-research.
- The user wants to build pivot tables, clean CSV data, or perform spreadsheet transformations — use
spreadsheet-analysis.
- The request is about investment portfolio optimization or securities analysis beyond basic position tracking.
- The user needs tax preparation or tax-code interpretation — flag as out-of-scope and recommend a qualified professional.
- The task is about building a financial software application — use coding-focused skills.
Operating procedure
Step 1 — Identify the financial artifact type
Determine which artifact the task requires:
| Artifact | When to use | Key principle |
|---|
| Ledger entry | Recording a transaction | Double-entry: debits = credits |
| Cashflow forecast | Projecting future cash position | 13-week rolling with weekly granularity |
| Debt schedule | Tracking loan repayment | Amortization with interest/principal split |
| Budget vs. actual | Comparing planned to real spending | Variance analysis with materiality threshold |
| Reconciliation | Matching two record sources | Every discrepancy must be classified |
Step 2 — Apply double-entry bookkeeping principles
Every transaction must follow the fundamental accounting equation:
Assets = Liabilities + Equity
For every ledger entry:
- Identify the accounts affected (at minimum two).
- Determine which accounts are debited and which are credited.
- Verify that total debits = total credits for the entry.
- Classify each account: Asset, Liability, Equity, Revenue, or Expense.
- Record date, description, reference number, and amounts.
Standard entry format:
Date: YYYY-MM-DD
Reference: [invoice/receipt/memo number]
Description: [what happened and why]
Debit: [Account Name] $X,XXX.XX
Credit: [Account Name] $X,XXX.XX
If debits ≠ credits, the entry is rejected — do not proceed until balanced.
Step 3 — Cashflow projection (13-week rolling forecast)
Build the forecast with weekly granularity:
- Opening balance: Start with confirmed bank balance as of the most recent reconciliation date.
- Cash inflows: List expected receipts by week — categorize as Confirmed (contract/invoice), Probable (>75% likelihood), or Possible (<75%). Only Confirmed and Probable enter the base case.
- Cash outflows: List committed payments (payroll, rent, loan payments, subscriptions) and discretionary spending by week.
- Net cash flow: Inflows − Outflows per week.
- Closing balance: Opening + Net cash flow = Closing. This week's closing = next week's opening.
- Runway calculation: At current burn rate, how many weeks until cash reaches zero or minimum threshold?
Update the forecast weekly: move actuals into the historical column, extend the projection window by one week, and adjust estimates based on new information.
Step 4 — Debt tracking and payoff strategy
For each debt instrument, build an amortization schedule:
| Payment # | Date | Payment | Principal | Interest | Remaining balance |
|---|
Compare payoff strategies when multiple debts exist:
- Avalanche method: Pay minimums on all debts; direct extra payments to the highest-interest-rate debt first. Minimizes total interest paid.
- Snowball method: Pay minimums on all debts; direct extra payments to the smallest-balance debt first. Maximizes psychological momentum via quick wins.
Present both strategies with: total interest paid, total time to payoff, and monthly payment requirements. Recommend avalanche unless the user has explicitly stated a preference for snowball's behavioral benefits.
Step 5 — Budget vs. actual variance analysis
For each budget line item:
- Record budgeted amount and actual amount.
- Calculate variance (Actual − Budget) and variance % ((Actual − Budget) / Budget × 100).
- Classify each variance:
- Favorable: Actual spending below budget (expenses) or actual revenue above budget (income).
- Unfavorable: The reverse.
- Apply materiality threshold: Only investigate variances exceeding 5% AND $500 (or user-specified thresholds). Immaterial variances are noted but not analyzed.
- For material variances, provide a root cause (price variance, volume variance, timing variance, or one-time event) and a corrective action if unfavorable.
Step 6 — Reconciliation
Match ledger entries against an external source (bank statement, invoice register, etc.):
- Match confirmed items: Entries that appear in both records with matching dates and amounts.
- Classify discrepancies:
- Timing differences: Transaction recorded in one period in the ledger, another in the bank (e.g., outstanding checks). Expected to clear; no action needed beyond tracking.
- Recording errors: Wrong amount, wrong account, duplicate entry. Requires correcting journal entry.
- Unrecorded transactions: Bank fees, interest, or automatic payments not yet in the ledger. Requires new journal entries.
- Unexplained differences: Cannot be classified. Escalate for investigation.
- Reconciliation status: RECONCILED (all items matched or explained) or UNRECONCILED (unexplained differences remain; list them).
Decision rules
- Debits must equal credits. No exceptions. Reject any entry that doesn't balance.
- Never mix personal and business transactions in the same ledger or account. If encountered, flag immediately and recommend separation.
- Forecasts must state assumptions. Every cashflow projection number must trace to either a confirmed commitment or a stated assumption with a probability estimate.
- Use the most conservative reasonable estimate for cashflow forecasts: underestimate inflows, overestimate outflows. Optimistic forecasts that prove wrong cause more damage than conservative ones.
- Materiality thresholds are mandatory for variance analysis. Investigating every $5 variance wastes effort; define thresholds upfront.
- Reconciliation must happen before reporting. Never produce financial summaries from unreconciled ledgers. State reconciliation status prominently.
- Debt payoff recommendations default to avalanche unless the user explicitly requests snowball or states behavioral preference.
Output structure
Every financial deliverable must use the appropriate format:
Ledger Entry
## Journal Entry: [Reference #]
- Date: YYYY-MM-DD
- Description: [transaction description]
- Accounts:
| Account | Debit | Credit |
|------------------|------------|------------|
| [Account Name] | $X,XXX.XX | |
| [Account Name] | | $X,XXX.XX |
- Balance check: Debits $X,XXX.XX = Credits $X,XXX.XX ✓
- Notes: [any additional context]
Cashflow Forecast
## 13-Week Cashflow Forecast (as of YYYY-MM-DD)
- Opening balance: $XX,XXX
- Assumptions: [list key assumptions]
- Forecast:
| Week | Inflows | Outflows | Net | Closing Balance |
|------|---------|----------|-----|-----------------|
- Runway: XX weeks at current burn rate
- Risk scenarios: [best case / base case / worst case closing balance at week 13]
Variance Report
## Budget vs. Actual: [Period]
- Materiality threshold: X% and $XXX
- Summary:
| Category | Budget | Actual | Variance | Var % | Status |
|----------|--------|--------|----------|-------|--------|
- Material variances requiring investigation:
- [Category]: Root cause, corrective action
- Overall assessment: ON TRACK / CAUTION / OFF TRACK
Reconciliation Status
## Reconciliation: [Account] as of YYYY-MM-DD
- Ledger balance: $XX,XXX.XX
- External balance: $XX,XXX.XX
- Difference: $XX.XX
- Classified items:
- Timing differences: $XX.XX (X items)
- Recording errors: $XX.XX (X items — correcting entries attached)
- Unrecorded transactions: $XX.XX (X items — new entries attached)
- Unexplained: $XX.XX (X items — ESCALATE)
- Status: RECONCILED / UNRECONCILED
Anti-patterns
- Mixing personal and business finances: Commingling funds in a single ledger or account makes reconciliation unreliable and creates legal/tax exposure. Always separate.
- Unreconciled balances treated as accurate: Reporting from a ledger that hasn't been reconciled against bank statements. The numbers may be wrong. Reconcile first.
- Forecasts without stated assumptions: A cashflow projection that says "we'll receive $50K in Week 3" without stating whether that's confirmed revenue or an estimate. Every figure needs a basis.
- Single-entry bookkeeping: Recording only one side of a transaction (e.g., "paid $500 for supplies" without crediting the cash account). This breaks the accounting equation and makes error detection impossible.
- Ignoring materiality: Spending hours investigating a $12 variance while a $5,000 discrepancy sits unexamined. Set thresholds and enforce them.
- Snowball-by-default: Recommending the snowball method without disclosing that it costs more in total interest. Always present both options with total-cost comparison.
Related skills
spreadsheet-analysis — Transforming raw financial CSV/Excel data into analysis-ready formats before ledger entry.
market-research — Providing revenue assumptions and market context that feed into cashflow forecasts.
business-idea-evaluation — Using financial projections as input to go/no-go business decisions.
competitor-teardown — Analyzing competitor pricing and margin structures to inform budget assumptions.
Failure handling
- If the ledger doesn't balance, stop all downstream analysis (forecasts, variance reports) until the imbalance is resolved. Report the discrepancy amount and affected accounts.
- If reconciliation reveals unexplained differences exceeding the materiality threshold, escalate before producing any summary reports. Do not smooth over the gap.
- If the user provides incomplete transaction data (e.g., amount but no accounts), ask for the missing information rather than guessing. Incorrect account classification corrupts all downstream reporting.
- If the task requires tax advice, regulatory interpretation, or investment recommendations, state that this is outside the skill's scope and recommend consulting a qualified professional.