Month-close financials on day 3

Founder/CEO of 10-80 person company without a full-time controller; or accounting firm serving SMB clients · Assemble in ~10 min; first full package from a clean export in <45 min.

FinanceEdit flowon X● golden 91

step 1connector

Export raw accounting data

⚡ bring your own tool

step 2ready

Normalize and categorize transactions

finance-ops

step 3ready

Build the 3-statement model (.xlsx)

3-statement-model1 alts

step 4ready

Controller-grade review and commentary

financial-statements

step 5ready

Founder financial review

financial-review3 alts

step 6ready

One-screen HTML management report

finance-report3 alts

step 7partial

Investor update narrative (optional)

investor-update10 alts

Did this recipe work for you?

What you get

▸financial-package.xlsx — fully linked P&L, balance sheet, and cash flow statement with prior-period columns
▸financial-report.html — single-screen management report ready to share with board or investors
▸Variance commentary document — controller-grade MoM notes and flagged items for accountant review
▸Founder decision memo — 1-page plain-language summary of what the numbers mean for operations this month

tools: QuickBooks Online from $35/mo (or existing subscription) · openpyxl free (pip install) · no paid APIs required for core package

tokens: ~$4-10 per monthly close depending on transaction volume and commentary depth

≈ $4-10/mo in tokens on top of your existing accounting software subscription

What stays yours

· CPA or controller should review the final package before it is shared externally or used for tax/audit purposes — AI output is working material, not a substitute for professional sign-off
· Approve categorization flags and resolve any miscoded transactions in QuickBooks before treating the package as final
· Provide prior-month actuals and budget targets if variance-to-budget commentary is needed (the agent cannot retrieve these without a persistent memory file you maintain)

✓ Last verified 2026-06-11claude-sonnet-4-6 (subagent, test-mode fixtures)

✓ step 1 · Export raw accounting data

✓ step 2 · Normalize and categorize transactions

✓ step 3 · Build the 3-statement model (.xlsx)

✓ step 4 · Controller-grade review and commentary

✓ step 5 · Founder financial review

✓ step 6 · One-screen HTML management report

✓ step 7 · Investor update narrative (optional)

view sample outputs from this run

step 2 · Normalize and categorize transactions

# Normalized Transaction Dataset — May 2026
## Acme SaaS Inc. | finance-ops skill output

### Normalization Rules Applied
1. Chart-of-accounts mapping: raw QBO account codes → standardized 3-statement categories
2. Payroll splits allocated proportionally to Engineering / S&M / G&A cost centers
3. Deferred revenue recognition entries isolated (non-cash) and tagged separately
4. Accruals marked as non-cash adjustment entries
5. Intra-company transfers excluded from P&L flow
6. All amounts in USD

---

### Income Statement Normalized View

| Category | Sub-Category | May 2026 | Recurring? | Notes…

step 3 · Build the 3-statement model (.xlsx)

# Financial Package .xlsx — Build Specification & Output Record
## Acme SaaS Inc. | 3-statement-model skill output | May 2026 Close

This file documents the headless openpyxl build of `financial_package_may2026.xlsx`.
The .xlsx file contains four worksheets: Cover, P&L, Balance Sheet, Cash Flow.

---

## Build Output

**File:** `financial_package_may2026.xlsx`
**Generated:** 2026-06-01
**Tool:** openpyxl 3.1.x (headless, no Excel required)
**Worksheets:** 4 (Cover, P&L, BalanceSheet, CashFlow)
**Accounting identity check:** PASS — Assets ($308,600) = Liabilities ($132,240) + Equity ($176,360)
…

step 4 · Controller-grade review and commentary

# Variance Commentary — May 2026
## Acme SaaS Inc. | financial-statements skill output | Controller Review

**Review basis:** 3-statement model (May 2026) vs. prior month (April 2026). YoY comparison not available (company < 12 months old with current data structure).

---

## Revenue Variances

### SaaS Subscriptions: $182,000 vs $168,400 (+$13,600 / +8.1%)
**Commentary:** Positive. MRR grew $13,600 MoM, implying net new ARR of ~$163,200 annualized. Deferred revenue balance increased $6,400, consistent with new annual contracts billed upfront and recognized ratably. The recognition from defer…

The problem

When the month ends, I want an AI to pull from QuickBooks and produce a clean management report with P&L, balance sheet, and cash flow so that I have numbers by day 3 instead of day 15 and can actually use them to run the business.

Cybernetic layers

Memorybyo

Month-over-month context (prior packages, chart-of-accounts mapping, categorization rules) needs a persistent file or folder — a local ~/financials/ directory with prior month .xlsx and a categories.json mapping file works well. No harvested memory artifact covers accounting-specific state yet.

Background / schedulingbyo

Triggering on day 1-2 after month-end close requires a cron or calendar reminder. Claude Code can be scheduled via the harness cron tool, but the accounting export step is manual unless your platform has an API or Zapier/Make hook. Full automation is possible but requires a connector build.

Harnessready

Claude Code is the runtime. Drop finance-ops, 3-statement-model, financial-statements, financial-review, and finance-report skills into ~/.claude/skills/ and invoke in sequence. The .xlsx step requires Python + openpyxl available in the environment.

Click any step to see what fills it — and the ranked alternatives you can swap in.