← ClaudeAtlas

datatablelisted

Renders tabular data as an interactive HTML table (sort, per-column filter, search, CSV export, hide columns, per-row annotation/labeling) opened in the browser. Use when the user says 'datatable', 'data table', 'show as a table', 'make this table interactive', or wants to visually explore, filter, or label a pandas DataFrame, a SQL/Spark result, or a CSV/TSV/Excel/JSON/Parquet file. A bare 'datatable this' means render the data already in scope. Not for web pages, charts, or non-tabular UI.
spunt/datatable-skill · ★ 1 · Data & Documents · score 64
Install: claude install-skill spunt/datatable-skill
# datatable skill Turn a DataFrame or a tabular file into an interactive HTML table and open it in the browser. The bundled script is `scripts/df2datatable.py`, located **inside this skill's own directory** — wherever this skill is installed (`~/.agents/...`, a plugin cache, a cloned repo, a shared team dir). Throughout this doc, `$SKILL_DIR` means the directory containing this SKILL.md; resolve it to the real absolute path the skill was loaded from and substitute it into every command below. Do not copy a fixed location. **Inputs:** CSV, TSV, Excel (`.xlsx`/`.xls`/`.xlsm`), JSON, Parquet. Unknown extensions — including `.txt` — are read as CSV. **Network requirement:** the output is a single HTML file, but it is **not offline-capable**. jQuery and the DataTables bundles load from `code.jquery.com` and `cdn.datatables.net` when the page opens. Without network access the page degrades to a plain unstyled HTML table — no sorting, filtering, export, or annotation. Flag this when the viewer is air-gapped, on a locked-down network, or the file is being emailed for later offline reading. **Opening:** on macOS the file auto-opens via `open`. Elsewhere that command either does not exist or means something else, so it fails (possibly printing a shell error) — the file is already written by that point, so hand the user the printed path. Pass `--no-open` to skip the attempt entirely. ## When to Use - Rendering a DataFrame, query result, or data file for visual exploration - Ad-ho