---
name: metadist-csv-insights
description: "Profile a CSV file and produce a Markdown report with row/column counts, fill rates, unique values, and numeric or categorical summaries."
license: "Apache-2.0"
metadata:
  decimalai-skill-id: "112fc236-c414-432b-aeb8-ae13f93fc8c4"
  decimalai-version-id: "fe231e1c-4ede-4fc1-830c-66dc2bfa3d8e"
  decimalai-version: "1"
  decimalai-body-sha256: "d2706b08a55b8fe0a0c34641a393b2f691e5f2bf958c9181e0dac6543607e906"
  decimalai-name: "metadist/csv-insights"
  decimalai-format: "1"
  decimalai-invocation: "model"
  decimalai-category: "data"
  decimalai-source: "https://app.decimal.ai/s/metadist-csv-insights@1/SKILL.md"
---
# CSV insights

Turn a CSV file into a readable Markdown report: total rows and columns, and for
each column its fill rate, unique values, and either numeric stats
(min/max/mean/std) or the most common categories.

## How to run

Run with Python. Pass the input CSV and an output Markdown path inside the
out-box:

```
python3 <skill_dir>/run.py <input.csv> <output.md> [title]
```

- `<input.csv>` — a CSV in a folder the user allowed, or already in the out-box.
- `<output.md>` — a path **inside the out-box**.
- `[title]` — optional report title.

Standard library only; no packages to install. It reads one CSV and writes one
Markdown file. After running, tell the user the saved path.