data-tools

Data Tools Skill

Critical Rule

NEVER use grep, sed, awk, or inline interpreters (python3 -c, node -e) on JSON, JSONL, YAML, TOML, XML, or CSV. Refuse and use the correct tool — text tools break on structure; inline scripts are verbose and fragile.

---

Key Patterns

For the general legacy-tool → modern-tool comparison (grep→rg, find→fd, cat→bat, ...), see the cli-tools skill's table.

Convert: dasel -w FORMAT or yq -o FORMAT.

jq -- JSON (no in-place)

jq -r '.version' package.json
jq '.users[] | select(.role == "admin") | .name' users.json
jq '.version = "2.0.0"' pkg.json > pkg.json.tmp && mv pkg.json.tmp pkg.json
jq -s '.[0] * .[1]' base.json override.json

Or gh --jq when the source is the GitHub CLI (see below) — no separate jq process needed.

yq -- YAML (in-place -i)

yq '.services.web.image' docker-compose.yml
yq -i '.services.app.image = "node:20"' docker-compose.yml
yq -o json config.yml

dasel -- TOML (only native tool) / XML (or xmlstarlet for XPath) / Universal auto-detect

dasel -f Cargo.toml '.package.version'
dasel put -f config.toml -t string -v "2.0" '.project.version'
dasel -f input.json -w yaml

qsv -- CSV / TSV

qsv headers data.csv && qsv stats data.csv --everything | qsv table
qsv search -s status "active" users.csv | qsv select name,email
qsv join user_id orders.csv id users.csv

qsv index, stats and frequency write files next to the input (x.csv.idx, x.stats.csv*). In a repository or a package directory, run them on a copy (details).

mlr -- JSONL / multi-format / DSL (in-place -I)

mlr --c2j cat data.csv
mlr --jsonl filter '$level == "error"' logs.jsonl
mlr -I --csv put '$total = $qty * $price' orders.csv

GitHub CLI -- always --jq

gh api repos/owner/repo/releases/latest --jq '.tag_name'
gh pr list --json number,title --jq '.[] | [.number, .title] | @tsv'

---

Anti-Patterns

# BAD → GOOD: jq -r '.version' package.json
grep '"version"' package.json | sed 's/.*"\(.*\)".*/\1/'
# BAD → GOOD: yq -i '.services.app.image = "node:20"' compose.yml
sed -i 's/image: node:.*/image: node:20/' docker-compose.yml
# BAD → GOOD: qsv select 2 data.csv
awk -F',' '{print $2}' data.csv
# BAD → GOOD: jq -r '.version' plugin.json
python3 -c 'import json;print(json.load(open("plugin.json"))["version"])'

---

References

Cookbook	Content
jq Cookbook	Filtering, transforms, GitHub CLI, large inputs ("Argument list too long", Windows)
yq Cookbook	Actions, Compose, K8s
dasel Cookbook	TOML/XML, conversion
CSV Processing	qsv workflows, joins
mlr Cookbook	JSONL, DSL, stats, joins
Enforcement Hook	PreToolUse gate shipped with this plugin: denies field extraction, warns once otherwise

Docs: jq, yq, dasel, qsv, mlr.