Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/skillmedev/document-studio/spreadsheet-model-buildernpx skills add SkillMedev/document-studio --skill spreadsheet-model-buildergit clone --depth 1 https://github.com/SkillMedev/document-studioWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00142 | $0.01666 |
| Opus 5 | $0.00071 | $0.00833 |
| Sonnet 5 | $0.00028 | $0.00333 |
| Haiku 4.5 | $0.00014 | $0.00167 |
Grade A, and why
Spreadsheet Model Builder scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 108 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Spreadsheet Model Builder
A spreadsheet model is judged by whether a second analyst can verify it without a guided tour. The costly failure is not a wrong formula - it is a hidden wrong formula: a hardcoded number buried in cell F47, an inconsistent row that Ctrl+\ would have caught, a scenario someone ran by overtyping inputs and never restoring. This skill builds models where errors have nowhere to hide.
Operating procedure
Structure comes before formulas: retrofitting sheet separation onto a finished tangle costs more than the model did.
Step 1: gather inputs
- What question the model answers and who reviews it. A model only the builder reads can cut corners; a reviewed model cannot.
- The time axis: granularity (monthly/quarterly), horizon, and history vs forecast boundary.
- Every assumption the user already holds, each with a source. Anything unsourced gets written down anyway - labeled a guess in the note column, never silently blended with sourced numbers.
- How many scenarios are needed. Decide now: each scenario is a column on the Inputs sheet, never a separate file.
Step 2: lay out the sheet structure
Separate sheets with single roles:
- Inputs (or Assumptions) - every editable variable lives here, and only here.
- Calcs - derived values only; no raw inputs, nothing typed by hand.
- Output (or Summary) - the numbers a reader cares about, linked from Calcs.
- Data - raw imported or pasted data, never edited by hand.
- Cover - what the model does, key assumptions, what changed in the last update.
The iron rule: no hardcoded number inside a formula, anywhere. If a value might ever change, it is an input. =D12*1.05 is a bug even when 5% is correct today, because nobody will find it when it stops being correct.
Step 3: build the Inputs sheet
- Group assumptions by theme: revenue, costs, headcount, timing.
- Every row carries a clear label, units, and the basis for the number in an adjacent note column ("per signed MSA", "mgmt guess", "trailing 6-mo actual").
- Color-code inputs distinctly (blue text on white is the convention) so an auditor identifies editable cells at a glance without a legend.
- Date the assumption set on the sheet.
- Scenarios (base / upside / downside) are side-by-side columns with one live selector cell; formulas read the selected column via INDEX, so switching scenarios is one edit, and no one ever overtypes the base case.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- yesterday First seen · 108 lines · 142 tokens per session scan A a87585141bbc
Spreadsheet Model Builder is a skill published in the GitHub repository SkillMedev/document-studio (1 stars, last pushed 1mo ago), licensed MIT. It adds 142 tokens to every session and 1,666 once invoked, about $0.0007 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
markdown-documents
Use when creating or editing Markdown documents, notes, reports, briefs, drafts, or other editable writing where Markdown should be the primary artifact format.
writer-documents
Use when creating, opening, or editing LibreOffice Writer ODT documents, or DOCX documents only when Microsoft Word compatibility is explicitly required.
pipefy-reports
Use this skill when the user wants to create, read, update, delete, or export pipe reports or organization reports. Covers the async export workflow (trigger, poll, download). 17 MCP tools.
html-artifact
Create self-contained HTML artifacts (single-file by default, split bundles when oversized) for complex, reviewable, or shareable work outputs. Use for an HTML artifact, browser-viewable report, implementation plan, PR/code-review report, dashboard, or mini deck, or for long Markdown-like outputs needing navigation or…
making-academic-presentations
Create academic presentation slide decks and optionally demo videos from research papers.
erpclaw
AI-native ERP system. Full accounting, invoicing, inventory, purchasing, tax, billing, HR, payroll, advanced accounting (ASC 606/842, intercompany, consolidation), and financial reporting (including P&L / trial balance / spend grouped by department, project, cost center, location, or fund). 496 actions across 14…