witan-xlsx-excelscript

npx skills add witanlabs/witan-cli \
  --skill witan-xlsx-excelscript

Running in Claude Cowork? The witan CLI isn't preinstalled — see references/cowork-setup.md for install steps.

Goal

Two jobs, one tool:

• Read & Analyze an existing workbook — find data, trace how figures are calculated, answer what-if questions — without guessing or corrupting it.
• Produce & Edit workbooks that are correct, polished, and idiomatic: formula-driven, sensibly formatted, and matched to the workbook's domain.

You are judged on correctness, layout, readability, and idiomatic style. A workbook with formula errors is not finished.

The tool

witan xlsx exec runs sandboxed JavaScript against a workbook opened server-side; this skill drives it with the Office Scripts (ExcelScript) dialect. Siblings calc, lint, and render handle verification and previews (below). Invoke with --stdin and a single-quoted heredoc (safe for every sheet name, no shell escaping):

witan xlsx exec report.xlsx --stdin <<'JS'
// @office-script
function main(workbook) {
  const sheets = workbook.getWorksheets().map(s => s.getName());
  const used = workbook.getWorksheet("Summary").getUsedRange();
  return { sheets, address: used.getAddress(), values: used.getValues() };
}
JS

• // @office-script MUST be the first line — it selects the ExcelScript dialect. Omit it and the script doesn't error: it silently runs in the other dialect, never calls main, and returns null. Always lead with it.
• function main(workbook) is the entry point — required. The engine calls it with the workbook and returns whatever you return (must be JSON-serializable). With --input-json, the parsed value arrives as a second arg: function main(workbook, input).
• ExcelScript is synchronous — getValue() after setFormula returns the recalculated value directly (no await needed); read results this way, don't recompute in JS. No imports.
• --create makes a new workbook (path need not exist; .xlsx only). It starts empty — call workbook.addWorksheet(name) before getActiveWorksheet().
• --save persists changes. Without it every write is ephemeral — it applies in the server session, recalculates, then is discarded; so reads and what-ifs never risk the file, and each run starts clean from the original.

After reading this file, you MUST read references/api.md before your first witan xlsx exec call — the ExcelScript surface is large and not guessable.

api.md has the CLI flags, the function main(workbook) contract, and worked recipes you can grep — reading, what-if, conditional formatting, charts, images, tables — plus what ExcelScript can't do, then points on to the full type surface in references/excelscript.d.ts.

Work efficiently (latency matters)

• Batch independent reads/writes into one exec; main takes a whole script, so prefer one rich call over many small ones.
• Don't re-read what you already pulled earlier in the task — reuse it.
• Exception — what-if: there, deliberately split into separate exec calls so you can review what you found before editing (see below).

Quality floor

Applies to every workbook you create or change. When editing an existing workbook, inspect or render it first and match its established conventions — the rules below are defaults for new or unspecified work, never a licence to restyle someone's model.

• Formulas, not values — every derived number is a formula so the sheet stays live.
  • ✅ sheet.getRange("B10").setFormula("=SUM(B2:B9)")
  • ❌ summing in JS and writing sheet.getRange("B10").setValue(4200)
• No magic numbers — reference an input cell, don't bury a constant.
  • ✅ =B5*(1+$B$6) ❌ =B5*1.05
• Separate inputs from logic — assumptions (rates, growth, multiples) live in their own labelled cells or an Inputs area, never inside formulas.
• Right references — absolute ($B$6) vs relative so fills and copies behave.
• Format to the data type — real dates with a date format (yyyy-mm-dd), not date-looking text; thousands separators for money/counts; sensible decimals on percentages; state units in the header when ambiguous (Revenue ($000s)).
• Lay it out for a human — styled header row; numbers right-aligned, labels left; sane column widths (cap autofit so nothing runs off-screen); modest row heights; whitespace between sections.
• Stay valid — prefix literal text starting with = using ' so it isn't read as a formula.

Domain conventions go beyond this floor. For financial models, valuations, projections, or IB work, read references/domains/financial-modelling.md.

Reading & understanding a workbook

• Get the lay of the land — workbook.getWorksheets() for the sheet list (.map(s => s.getName())); sheet.getUsedRange() for the populated region, then .getValues() to dump it cheaply (or .getFormulas() to see the logic). getUsedRange(true) trims to value-bearing cells.
• Find things — range.find(text, { completeMatch, matchCase }) returns the matching cell, or a nullish value if absent. Matching is exact/substring, not fuzzy or synonym-aware — search the literal label, or pull getValues() and scan in JS. Tables come from workbook.getTables() (table.getColumnByName(name)).
• Disambiguate — a label and the formula cell beside it often share text. Read surrounding cells with getValues()/getFormulas(), then pick the formula cell (the one whose getFormula() starts with =), not the label.
• Trace calculations — ExcelScript has no dependency-tracing API (no precedents/dependents). Trace by hand: read the target cell's getFormula(), follow each cell reference it names, and repeat. Filter to the formulas you care about; don't dump whole sheets.

What-if / sensitivity

For "what happens to Y if X changes?", use two separate exec calls so you verify before editing:

1. Locate the output cell (first call) — search for what's asked about, read its getFormula(), confirm you have the formula cell.
2. Change + read (second call) — read the baseline getValue(), setValue on input X, then read Y's getValue() again; recalc is automatic, so the new number is there immediately. If Y didn't move, X doesn't drive it ⇒ wrong cell.
3. Report baseline → new, and re-read a few downstream cells (or run witan xlsx calc) to confirm your edit broke nothing.

There is no sweepInputs helper — for many combinations, loop inside one main: set each input, read the outputs, collect the before/after rows, and return them. Circular/iterative models: a standalone full-workbook pass is witan xlsx calc; non-convergence shows up there as errors.

Authoring a workbook

Build in this order — it minimises rework and round-trips:

1. Plan the structure — sheets, and where inputs vs calculations vs outputs live. Inputs first.
2. Write content in bulk — headers, data, and formulas in as few calls as possible. Prefer setValues/setFormulas with a 2-D array over a region to writing each cell. Gotcha: unlike desktop Excel, setFormula on a multi-cell range writes the same literal formula to every cell — it does not adjust references relatively (B1:B3 ← "=A1*2" makes all three =A1*2). To fill a relative formula down, build the per-cell strings and pass them to setFormulas ([["=A1*2"],["=A2*2"],["=A3*2"]]).
3. Format — number formats (setNumberFormat), header styling (getFormat().getFont()/getFill()), widths/heights, alignment, borders (getFormat().getRangeBorder(ExcelScript.BorderIndex.edgeTop)). Quality floor above; finance models also follow the domain reference.
4. Add interactivity as needed — data validation for categorical inputs, conditional formatting, then charts/tables.
5. Verify (below) before calling it done.

Prototype ephemerally (omit --save) while iterating; add --save only once the structure is right.

Verify before done — mandatory

A workbook with formula errors is not delivered. Before finishing any authoring or edit:

• Recalculate and check errors — witan xlsx calc model.xlsx must report 0 errors. It prints every error with address and formula, and exits non-zero. Fix and repeat until clean.
• Spot-check key ranges — read formulas back (getFormulas) on totals, a few sample references, and edge rows.
• Confirm layout — witan xlsx render on the headline range; headers, merges, number formats, and charts look as intended.
• Lint for logic errors — witan xlsx lint model.xlsx flags what calc can't: double-counting, approximate-match lookups on unsorted data, mixed currencies/units. Review and resolve or knowingly accept each finding; exits 2 on any.

Then report what you built (sheets / ranges), and for what-ifs the baseline → new values.