What problem does it solve? Building financial models in Excel by hand leads to silent errors: hardcoded values where formulas should be, undocumented assumptions, and no way for a reviewer to audit the work. This Skill produces .xlsx files headlessly via openpyxl that follow strict banker-grade conventions so any reviewer can immediately distinguish inputs from calculations. ## Core Features & Use Cases - Auditable color conventions: Blue fonts for hardcoded inputs, black for live formulas, green for cross-sheet links, plus cell comments documenting the source of every hardcoded number. - Formula-first modeling: Every derived cell is written as a live Excel formula, with named ranges for cross-sheet references and a dedicated Checks tab that surfaces TRUE/FALSE balance validations. - Sensitivity tables and recalculation: Builds odd-sized sensitivity grids with a highlighted base-case center cell, and includes a script to recalculate formulas via headless LibreOffice before delivery. - Use Case: Ask the agent to build a DCF model for a company. It creates an Inputs tab with sourced assumptions, a DCF tab of live formulas, a Checks tab verifying the model ties out, and a 5x5 WACC/terminal-growth sensitivity table, saved to ./out/model.xlsx. ## Quick Start Use the excel-author skill to build a DCF valuation workbook for Acme Corp with sourced inputs, live formulas, a checks tab, and a WACC sensitivity table saved to ./out/acme-dcf.xlsx.