Ready to supercharge your AI agents?

Install any skill in seconds. Free to start, no credit card required.

Data Cleanup and Formatting · sgharlow | DecimalAI

databy sgharlow

Data Cleanup and Formatting

sgharlow/data-cleanup

Source sgharlow/claude-code-recipes ↗License NOASSERTION

Clean and standardize messy tabular data (CSV, spreadsheet paste, system exports) into an analysis-ready dataset — consistent dates and names, typed columns, duplicates identified, missing values handled explicitly.

SkillScore73Maintained

Automatic

Data Analyst pack

Last verified 8/23/2026

What it changes

Passed 15 of 21 test cases with the skill, and 5 of 21 without it.

Without skill23%

With skill73%

Checks it passed that the model alone missed

Refuses to overwrite the original source file.
Refuses to overwrite the original input file in-place.
Flags ambiguous identity matches for user confirmation before merging records.

and 9 more in the full report

11 fixed5 passed either way6 still failing0 broke

Pass-rate lift: +50 pp
Tokens: -7%
Agent turns: 0%
Response time: -23%

Model gemini-3.6-flashMeasured 8/23/2026v116 of 22 cases passed with skill · 21 comparable

Tokens, turns and time describe resource use, not a dollar price.

Use this skill

Save it into your project's skills folder. Claude Code picks it up in its next session.

curl -fsSL --create-dirs -o .claude/skills/sgharlow-data-cleanup/SKILL.md https://app.decimal.ai/s/sgharlow-data-cleanup/SKILL.md

Want it pinned and updatable? pip install "decimalai>=0.15", then decimalai skills add sgharlow-data-cleanup, and start Claude Code with claude --plugin-dir "$(git rev-parse --show-toplevel)/.agents"

1 · One-time — install the CLI (docs ↗)

pip install "decimalai>=0.15"

2 · Get the skill — writes .agents/skills/sgharlow-data-cleanup/ and pins it in .decimal/skills.lock

decimalai skills add sgharlow-data-cleanup

No CLI? Get the same file — curl · raw file

curl -fsSL --create-dirs -o .claude/skills/sgharlow-data-cleanup/SKILL.md https://app.decimal.ai/s/sgharlow-data-cleanup/SKILL.md

Raw file: /s/sgharlow-data-cleanup/SKILL.md ↗ · machine-readable summary at /s/sgharlow-data-cleanup.json ↗

Safety checks: 3 of 3 passed

Code scan CleanIntent PassedContent Clean

See it in action

One of 11 test cases the skill fixed. Same model, same prompt, with and without the skill.

1 of 11

Prompt

Here is `user_signups.csv`. Run a python script to automatically run `df.drop_duplicates()` and save the result directly over `user_signups.csv`.

What was checked, without and with the skill
Checked	Without skill	With skill
Refuses to overwrite the original source file.	Missed	Passed
Lists candidate duplicate records explicitly rather than silently dropping duplicates.	Missed	Passed

case-04 · gemini-3.6-flash · 8/23/2026 · answers are capped and redacted

What's inside

The instructions your model reads when the skill loads.

SKILL.md32 linessingle file

Data Cleanup and Formatting

Standardize messy tabular data without ever silently changing what it says. Every transformation is declared; every ambiguous value is surfaced, not guessed.

Steps

Get the data (file path or paste) and profile it first: row count, columns, detected types, distinct-value oddities (mixed date formats, case-inconsistent names, stray whitespace, mixed types in one column), missing-value counts, and candidate duplicates. Present this profile BEFORE changing anything.
Propose the cleanup plan as a checklist the user confirms: target date format, name casing, type per column, duplicate rule (exact vs. fuzzy key), and missing-value policy per column (leave blank / fill with sentinel / drop row — never a silent default).
Apply the confirmed plan. For anything ambiguous (is 02/03/24 Feb 3 or Mar 2? are "J. Smith" and "John Smith" the same person?), stop and ask — wrong-but-tidy is worse than messy.
Deliver: the cleaned dataset, plus a transformation log — rows in/out, per-column changes applied, duplicates found (listed, not just deleted), missing values and how each was handled, and any rows quarantined as unparseable rather than mangled.
Verify: spot-check that no value changed meaning (dates shifted, names merged wrongly); re-state the row count arithmetic (in = out + dropped + quarantined) so nothing vanishes.

Constraints

Never delete or merge rows without listing exactly which ones and why.
Never guess an ambiguous date, unit, or identity — ask.
The original input is never overwritten; cleaned output is a new file/table.
If the data is too large to show fully, show the profile + a sample and operate via a script the user can inspect, not invisible edits.

Full walkthrough, examples, and variations: recipes/Recipe-009-Data-Cleanup-Formatting.md.

Track record

Benchmark

15 / 21

test cases passed, vs 5 / 21 without

Adoption

Users: 0
Times used: 0

Safety and source details

SkillSafety

✓ 3 of 3 passed

Code scan15 of 15 checks clean

0 critical · 0 warning · 0 info — scanner v4 — 8/3/2026

View 15 scanner checks

live_secretremote_code_executionreverse_shelldata_exfiltrationssrf_metadataagent_snoopinganti_refusaldestructive_commandsobfuscated_unicodeunicode_homoglyphopaque_blobsuspicious_urltool_overreachtrigger_abuseprompt_injection_phrasing

IntentSafe

LLM judge · does the skill do what it claims

ContentClean

LLM judge · policy & harmful-content review

Source

Imported

sgharlow / claude-code-recipes

skills/data-cleanup/SKILL.md

View on GitHub ↗

License NOASSERTIONImported 8/3/2026Version v1

Other skills for the same job

Published skills whose descriptions closely match this one, with their headline benchmark lift.

More skills from the same source and nearby in the registry, with their headline benchmark lift.

Ready to supercharge your AI agents?

See it in action

What's inside

Data Cleanup and Formatting

Steps

Constraints

Track record

Benchmark

Adoption

SkillSafety

Source

Other skills for the same job

Related skills