Individual skill
spreadsheet-qa
Answers business questions about the user's own spreadsheet or data export (CSV, TSV, XLSX from a CRM, Shopify, Stripe, QuickBooks, ad platforms, HR or payroll systems) correctly and auditably. Profiles every file first (grain, duplicate keys, embedded Grand Total rows, text-stored numbers, mixed currencies, UTC timestamps, Excel serial dates, hidden rows), states the metric definition used, computes every number in SQL or pandas instead of mental math, checks joins for fan-out, reconciles to a known total, and returns the query, filters, row counts, and rows behind the answer. Use this whenever the user says "analyze this spreadsheet", "how many...", "what's the total in this export", "which customers...", "pivot this", "revenue by month", "top 10", "what's our churn/MRR/AOV", or asks any question about their data, CSV, or Excel file, even a quick one. Also use when they ask to change values in a data file.
spreadsheet-qa/SKILL.md
Source commit b2267d9171af4f56848d82992381b20b89b44056
What was checked
- Matches the observed repository HEAD
- Yes
- Scanner rules are current
- Yes
- Repository scan complete
- No
- Skill scan complete
- No
- Observed behavior
- Local scripts
- Last checked
- Nov 10, 2026
These are observations at the recorded commit, not a guarantee of safety. Review the source before running it.
Install this skill
Choose an agent in your Stack or copy a command below. Commands fetch the pinned commit and copy only this skill directory. Existing destinations are refused. Copying a command does not install anything.
An install command is unavailable: current rules, complete repository and skill scans, an eligible behavior level and a permitted license are required.
Repository context
- Repository stars (not skill ratings)
- 425
- Repository last updated
- Oct 20, 2026
- License
- MIT
Observed agent integration: No specific integration evidence. Standard SKILL.md installation does not imply verified compatibility.
- agent-skills Repository topic
- ai-agents Repository topic
- anthropic Repository topic
- claude Repository topic
- claude-code Repository topic
- claude-skills Repository topic
- developer-tools Repository topic
- llm Repository topic
- prompt-engineering Repository topic
- sales-automation Repository topic
- data Derived category
- sales Derived category
- data-databases Derived category
Files and excerpt
File metadata and permitted, redacted excerpts come from the recorded scan. Source links are pinned to that commit.
Browse files
SKILL.mdcore · 8,616 bytesSKILL.mdcore · 8,616 bytesSKILL.mdcore · 8,616 bytesSKILL.mdcore · 8,616 bytesSKILL.mdcore · 8,616 bytesSKILL.mdcore · 8,616 bytesreferences/metric-definitions.mddoc · 7,035 bytesreferences/traps.mddoc · 7,140 bytesscripts/ask.pyruns · 14,136 bytesscripts/diff_edit.pyruns · 4,236 bytesscripts/profile.pyruns · 27,150 bytestests/build_fixtures.pyruns · 4,571 bytestests/fixtures/customers.csvasset · 340 bytestests/fixtures/fx_rates.csvasset · 52 bytestests/fixtures/orders.csvasset · 880 bytestests/fixtures/pipeline.xlsxasset · 5,174 bytestests/fixtures/refunds.csvasset · 177 bytestests/questions/q1_net_revenue.jsonasset · 3,881 bytestests/questions/q2_august_ny.jsonasset · 2,576 bytestests/questions/q3_naive_fanout.jsonasset · 2,965 bytestests/questions/q3_segment_revenue.jsonasset · 3,528 bytestests/questions/q4_top_customers.jsonasset · 3,899 bytestests/run_tests.pyruns · 10,564 bytes
Read excerpt
# Spreadsheet QA
Business exports lie quietly. A "Grand Total" row doubles every SUM. A customer listed twice multiplies their orders through a join. EUR and USD sit in one amount column. A UTC timestamp at 02:30 on Aug 1 is a July order in New York. None of these raise an error, so an answer computed by eye or by one quick query looks right and is wrong. This skill makes every number reproducible: profile, define, compute in code, reconcile, show the evidence.
Scope: questions about the data in the file. To combine or dedupe several files into one first, use `csv-excel-merger`. To check whether a financial model's formulas are sound, use `spreadsheet-model-auditor`.
## Workflow
### 1. Profile before answering anything
```bash
python3 scripts/profile.py path/to/*.csv path/to/book.xlsx --out sqa_out
```
