spreadsheets

Spreadsheets

Handle tabular data with exact values, minimal diffs, local privacy, atomic writes, and structural validation.

Invariants

1. Keep precision-sensitive amounts as strings and compute with decimal.Decimal or DuckDB DECIMAL(38, 18), never binary floats.
2. Touch only requested rows, columns, formulas, and formatting. Existing file conventions override house defaults.
3. For newly authored text tables, prefer TSV, UTF-8 without BOM, LF, one trailing newline, lowercase snake_case headers, ISO dates, . decimals, and - nulls.
4. Read unknown text tables with BOM-tolerant UTF-8; never write a BOM.
5. Write in place atomically through a sibling temporary file, validate it, then replace the target.
6. Escape external cells beginning with =, +, or @; a bare - null is exempt. Formula-prefix cells in trusted authored data are observations, not proof of injection.
7. Keep transaction, bank, exchange, and tax data local. Redact samples unless raw rows were explicitly requested.

Factual Profiling

Resolve helper paths from this SKILL.md. Profile unknown data before choosing a transformation tool:

uv run "<skill-dir>/scripts/profile.py" <file> --redact-samples

The JSON output has schema_version: 2. It reports structural facts, header quality, cardinality/statistics when qsv is available, frequency facts, formula-prefix cells, workbook metadata, and local tool availability. It contains no tool recommendations and does not infer identifiers from uniqueness. Choose the tool from the requested transformation, provenance, output format, and preservation requirements.

Use --external-data only when the cells came from an external or otherwise untrusted source and will be written to a formula-capable consumer. With that flag, formula-prefix cells affect status; without it, legitimate formulas such as =SUM(...) remain factual observations and do not fail the profile.

Tool Routing

Prefer qsv --cache-threshold 0 where supported. When qsv stdout must remain TSV, use -o out.tsv; stdout otherwise defaults to CSV.

Workflow

1. Inspect with peek.py --redact-samples; add profile.py when cardinality, formula prefixes, metadata, or available tooling matters. For a no-shape-change edit, save the peek JSON. For intentional row/schema changes, record the expected width and invariants.
2. Decide whether formula-prefix cells are dangerous from provenance and output context. Decide the smallest tool that preserves values and formatting. Avoid pandas unless necessary; if used, load every column as strings.
3. Apply the transformation atomically. For idempotent appends with legitimate duplicate rows, use multiset difference rather than set deduplication.
4. Validate:
   • unchanged shape: peek.py --strict --expect-like <before-report>;
   • changed shape: peek.py --strict --expect-columns <n> plus task-specific counts/keys;
   • authored house TSV: add --house;
   • formulas: uv run scripts/recalc.py <file.xlsx> and require success.
5. Report paths, row/column effects, validation, and any workbook features that could not be preserved.

For human output, lead with ### 📊 Spreadsheet — ✅ updated only after the write and required validation pass, or ### 📊 Spreadsheet — 🔎 inspected, no files written for read-only work. On required validation failure, use ### 📊 Spreadsheet — ⛔ not deliverable. Do not paste private profile JSON or decorate cells, headers, formulas, paths, commands, or diagnostics.

prb-finance

Never hand-edit generated .pool.tsv, .annual.tsv, or Markdown reports. After source edits, run just tsv-check, then just cli::write-changed. Cap private output to counts and file references unless raw rows were requested.

Completion requires the requested artifact, an intentional diff, atomic replacement where applicable, and structural plus domain validation evidence.