| name | csv-stats |
| description | Compute mean, median, min, max, count for every numeric column in a CSV file. Use when the user asks for descriptive statistics, summary stats, or "tell me about this CSV". |
| tags | ["csv","statistics","analysis"] |
CSV stats
Use this skill when the user wants descriptive statistics for a CSV file.
Steps
- Use
read_text(path=<csv>) to load the file.
- Parse the header row by splitting the first line on
,.
- For every subsequent row, split on
, and try to coerce each cell
to float. Cells that fail the coercion are treated as missing
for that column.
- For every column with at least one numeric value, compute:
count, mean, median, min, max. Use the standard
formulas:
- mean = sum(values) / count
- median = sorted middle value (or average of two middles for
even-length lists)
- Render a Markdown table with one row per numeric column. Headers:
column | count | mean | median | min | max. Round all floats to
3 decimal places.
- Write the Markdown table to
<csv stem>_stats.md next to the
input file using write_text. Include a top-level heading
# Statistics for <filename>.
Output contract
The user expects a <stem>_stats.md file written to disk. The
done(answer=...) summary should mention the path.
