agentleFS
PSU3D0PSU3D0/formualizer-plugin/skills/spreadsheet/SKILL.md

spreadsheet

Skill0 starsChanged 6 days agoApache-2.0View on GitHub

Fast Excel (.xlsx) formula recalculation, targeted cell reads, batch edits, and dependency tracing. Use to explain unexpected totals, trace precedents or dependents, test assumptions, fix stale results, and calculate workbooks created or edited with other tools.

12 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

7 rules. 0 are common, 3 are distinctive, and 4 are very specific to this project.

  • 0

    Common

    Most of the 10 edited versions we read keep these.

  • 3

    Distinctive

    This author's own choices. Worth a look.

  • 4

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Test assumptions or edit cells/formulas: run edit to apply atomic set_cells changes and recalculate in one step, then inspect affected …

    line 9
  2. Very specific:

    Calculate a final deliverable from other tools: when creating or modifying a formula-bearing .xlsx with another library, run …

    line 10
  3. Distinctive:

    Check the bundled engine:

    line 18
  4. Distinctive:

    Check cached formula results without writing, or write recalculated caches while preserving non-cache package bytes:

    line 71
  5. Very specific:

    Only an explicit --output argument writes a workbook.

    line 82
  6. Very specific:

    Read receipt coverage, completed, and error counts before reporting calculated results.

    line 84
  7. Distinctive:

    Verify comparative narrative claims (such as minimums, troughs, percentage shares, or counterfactuals) against all relevant periods and …

    line 86

“Common” is measured: kept by at least 6 in 10 of the 10 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Route tasks to the matching formualizer:spreadsheet operation:

  • Explain formulas, drivers, or unexpected totals (read-only): run trace (--roots ... --direction precedents --max-depth 2) to map formula dependencies and inspect (--cells ... or --range ...) to read formulas and stored cached values. Do not run source-recalculate on explanation or read-only requests unless the user asks to verify or repair stale caches, or stored caches are missing.
  • Test assumptions or edit cells/formulas: run edit to apply atomic set_cells changes and recalculate in one step, then inspect affected output cells and verify the receipt.
  • Calculate a final deliverable from other tools: when creating or modifying a formula-bearing .xlsx with another library, run source-recalculate --output /path/final.xlsx (or --check for read-only verification) after saving final changes and before returning the file.

inspect and trace never evaluate formulas or modify files. Stored formula caches carry freshness: "unverified". Reading or explaining a workbook does not authorize recalculation.

Run locally

Requires Node >=20. The engine is bundled; no install, network access, credentials, or local server is required. Locate this SKILL.md and set CLAUDE_SKILL_DIR to its parent directory.

Check the bundled engine:

node "${CLAUDE_SKILL_DIR}/scripts/xlsx.mjs" doctor

1. Inspect cells or ranges (read-only)

Pass --cells or --range directly (addresses must be sheet-qualified, such as Sheet1!A1 or 'My Sheet'!A1:D20):

node "${CLAUDE_SKILL_DIR}/scripts/xlsx.mjs" inspect /path/input.xlsx --cells "Sheet1!B1" "Sheet1!G12"
node "${CLAUDE_SKILL_DIR}/scripts/xlsx.mjs" inspect /path/input.xlsx --range "Sheet1!A1:M20" --limit 100

Or pass canonical JSON via --json '<JSON>' or a REQUEST.json file. Top-level fields are singular selection and optional page only (never selections or sheets):

{"selection":{"kind":"cells","addresses":["Sheet1!A1","Sheet1!B1"]}}
{"selection":{"kind":"range","address":"Sheet1!A1:D20"},"page":{"limit":50}}

2. Trace precedents or dependents (read-only)

Trace upstream inputs (precedents) or downstream formulas (dependents):

node "${CLAUDE_SKILL_DIR}/scripts/xlsx.mjs" trace /path/input.xlsx --roots "Sheet1!G12" --direction precedents --max-depth 2
node "${CLAUDE_SKILL_DIR}/scripts/xlsx.mjs" trace /path/input.xlsx --roots "Assumptions!B9" --direction dependents --max-depth 2

Or pass canonical JSON via --json '<JSON>' or REQUEST.json (including verbatim next_trace_request or next_request cursors). Top-level fields are roots plus optional direction, limits, details, and page:

{"roots":["Sheet1!G12"],"direction":"precedents","limits":{"max_depth":2}}

Traces are bounded (max_depth defaults to 0 if omitted) and do not guarantee the entire dependency graph. Follow returned next_trace_request or next_inspect_request payloads for deeper hops. See stateless-inspection.md.

3. Batch edit assumptions or formulas

Apply atomic changes and recalculate in one call via --operations-json or --operations /path/ops.json:

node "${CLAUDE_SKILL_DIR}/scripts/xlsx.mjs" edit /path/input.xlsx \
  --operations-json '{"ops":[{"kind":"set_cells","sheet_name":"Sheet1","cells":{"A1":{"kind":"value","value":30}}}],"recalculate":"auto"}' \
  --output /path/edited.xlsx

To replace an existing formula in set_cells, include "overwrite_formulas": true and {"kind":"formula","formula":"A1*3"} (without leading =). Set "details": true on the request for an exact cell diff. Rich editing can rewrite the workbook package and does not guarantee original-package preservation; see stateless-edits.md.

4. Recalculate original workbook bytes

Check cached formula results without writing, or write recalculated caches while preserving non-cache package bytes:

node "${CLAUDE_SKILL_DIR}/scripts/xlsx.mjs" source-recalculate /path/input.xlsx --check
node "${CLAUDE_SKILL_DIR}/scripts/xlsx.mjs" source-recalculate /path/input.xlsx --output /path/recalculated.xlsx

See source-runtime.md for --options / --options-json calculation settings and receipt fields.

Check and deliver

Only an explicit --output argument writes a workbook. It overwrites an existing destination, including the input if explicitly selected; parent directories must exist.

Read receipt coverage, completed, and error counts before reporting calculated results. bytesUnchanged and legacy current mean byte equality only. Report partial or deferred calculation explicitly; an available artifact alone does not establish completed calculation or fresh caches.

Verify comparative narrative claims (such as minimums, troughs, percentage shares, or counterfactuals) against all relevant periods and cells before stating them. Attach or link any requested saved workbook using the host's artifact mechanism and verify that the output file exists.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.