Tool Owl

What it does

When your agent uses it

• The user mentions Owl
• The owl CLI (owl new/run/snapshot/viz)
• Wants to create
• Debug a small dataset with Owl

Example prompts

• “/tool-owl”

Requirements

• Python 3

Workflow steps

7 steps, taken from the first numbered list in SKILL.md.

1. 1Scaffold the step (owl new, or hand-create vYYYYMMDD/), using a --date matching the source's access date.
2. 2Move the snapshot fetch into a @Snapshot (e.g. snap.download(URL, suffix=".csv")).
3. 3Collapse the meadow + garden transforms into one @Dataset — load with load_snapshot(...), harmonize, tb.format([...]).
4. 4Port the garden .meta.yml into the Owl meta.yml: the datasets. block accepts the standard catalog shape (definitions, tables → variables)…
5. 5If it was a grapher step, add an @Action(kind="grapher", default=False) that calls upsert_dataset(...).
6. 6Remove the classic files (snapshot .py/.dvc, meadow, garden + meta, grapher) and their active DAG entries.
7. 7owl snapshot + owl run, then verify the output dataset matches the old one.

What it can do on your machine

Read from SKILL.md and the folder at commit bf5dc8e. It shows what the files ask for, not the result of running them.

• Tool permissions

  Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

  From allowed-tools in the SKILL.md frontmatter.

• Runs code

  Shell commands in SKILL.md call:

  • make

  From the folder's file list and the shell code blocks in SKILL.md.

• Network

  No URLs in SKILL.md.

  From URLs in SKILL.md, links to its own repository left out.

• Credentials

  Names no API keys, tokens, secrets or passwords.

  From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Tool Owl loads about 1.4k tokens when it runs. Until then it costs about 85 tokens; SKILL.md has 496 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it

~85

When it runs · the whole SKILL.md, loaded when a task matches

~1.4k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from owid/etl at commit bf5dc8e, republished under its MIT licence (© owid). 496 words, ~1,350 tokens.

name

tool-owl

description

Working with Owl — OWID's lightweight, single-folder pipeline runner for small datasets (an alternative to the full snapshot → meadow → garden → grapher chain). Use whenever the user mentions Owl, `owl_steps/`, the `owl` CLI (`owl new`/`run`/`snapshot`/`viz`), or wants to create, migrate, run, or debug a small dataset with Owl.

metadata.internal

true

metadata.owner

Marigold

from pathlib import Path

from owid.catalog import Table
from owl import Action, Dataset, Snapshot, SnapshotCapture
from owl.catalog import load_snapshot
from owl.grapher import upsert_dataset

from etl.data_helpers import geo

URL_DOWNLOAD = "https://example.com/data.csv"
COUNTRIES_FILE = Path(__file__).with_name("my_dataset.countries.json")


@Snapshot
def raw_data(snap: SnapshotCapture) -> None:
    snap.download(URL_DOWNLOAD, suffix=".csv")


@Dataset
def my_dataset(raw_data: Snapshot) -> Table:
    tb = load_snapshot(raw_data)                       # Table with origins from meta.yml `origin`
    tb = tb.rename(columns={"country_name": "country"})
    tb = geo.harmonize_countries(df=tb, countries_file=COUNTRIES_FILE)
    tb = tb.format(["country", "year"])
    return tb


@Action(kind="grapher", default=False)
def upsert_to_grapher(my_dataset: Dataset) -> None:
    upsert_dataset(my_dataset)

© owid, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

Just SKILL.md in .claude/skills/tool-owl of owid/etl.

Open the folder on GitHubat commit bf5dc8e