Spreadsheets

Spreadsheets

For read-only questions, inspect the source and answer without exporting a replacement. For edits, preserve existing sheet structure, formulas, styles, validation, named ranges, charts and hidden content outside the requested change.

Environment and output

Use run_command with a verified Python: -c 'import openpyxl; import sys; print(sys.executable)'. If needed, check project .venv/bin/python or another configured Python environment. Use that absolute executable. Locate LibreOffice on PATH or in the selected runtime's dependencies/bin/override/soffice. Install missing dependencies only through current approval policy.

Resolve these references/scripts from the Skill directory returned by load_skill. Keep original inputs intact, write the final workbook inside the workspace, and use task-specific QA directories. Do not add worksheets containing implementation notes or test data to the delivered workbook.

Workflow

1. Read workbook authoring and verification. Inspect the source in both formula and cached-value modes. Identify each row's meaning, units, dates and identifiers before calculating anything. CSV/TSV require explicit delimiter/encoding and preservation of identifier strings.
2. Use openpyxl for supported XLSX edits, standard csv for delimited text. It writes formulas but does not calculate them. Use keep_vba=True for an XLSM copy and inspect macros/unsupported features before choosing a save path. Convert legacy XLS only on a copy with an available office engine.
3. Preserve source data and derive results with formulas. Use typed numeric/date values and number formats. Keep editable assumptions clearly labeled; use relative/mixed/absolute references intentionally. Do not hide missing inputs as plausible zero results.
4. Keep the workbook as small as the task permits. Put the main result above relevant detail; separate source and build sheets only when useful. Source → assumptions/build → output is the business dependency direction. Check/Audit calculations must not feed business results.
5. Save to a new file, then perform actual recalculation using the reference's LibreOffice procedure or the available target application. Reopen the recalculated file with data_only=True; setting fullCalcOnLoad is not proof of calculated values.
6. Run the supplied checker after recalculation:

   <python> <skill-directory>/scripts/check-workbook.py <recalculated.xlsx> --require-cached
   

   For independently known results, add --expect <expected.json> where JSON maps Sheet!Cell to expected values. Check representative first/middle/last rows, totals, input changes and boundary cases in a disposable copy. Confirm charts and statuses update with input changes.
7. Review every new or affected sheet visually. Use an office PDF preview and view_pdf_page; preserve source print settings and adjust only a QA copy if a screen-oriented sheet needs different print areas. Inspect for ####, clipped labels, tiny text, broken charts and inappropriate page splitting.
8. Deliver the final recalculated workbook with a file link. Preserve unsupported native features or report a concrete limitation. Do not claim native Excel behavior from formula text or cached values alone.

Advanced local operations

For advanced edits, template preservation and format-specific validation, read advanced workflow. Read only the sections relevant to the task.

Dependency setup

If the verified interpreter lacks required modules, create an isolated environment inside the workspace with python3 -m venv .artifacts/document-runtime and install this Skill's requirements.txt using that environment's python -m pip install -r <skill-directory>/requirements.txt, subject to the current command approval policy. Reuse the verified environment for subsequent calls. LibreOffice and Poppler are native dependencies: locate existing executables first; if absent, install through the platform package manager under the current policy and verify soffice --version and pdftoppm -v. Never treat missing dependencies as a successful render/recalculation.