Panel data steward

A Claude skill that finds messy participant data, like one job title stored four different ways, and proposes fixes that a person approves before anything changes.

Research opsOctober 12, 2026

Download the skill

Enter your details to download the skill file.

What it does

Participant data gets messy. The same job title is stored as "Product Manager", "product manager", "PM" and "Product Mgr", so filtering on one misses most of the people and the panel looks smaller than it is.

You give it an export of your participant panel. It tells you which fields are mostly empty, which are duplicates, and which values are different spellings of the same thing. It proposes fixes, and nothing changes until a person approves each one. Use it when your panel seems too small, or as a regular clean-up.

What you get back

  • An inventory of every attribute, how well it is filled in, and which names look like duplicates.
  • Proposed clusters of variant spellings, each with a canonical value to approve.
  • A review queue for close calls, and a list of merges it refused and why.
  • A per-person list of exactly what would change, before anyone approves anything.
  • A write plan with an audit log and a reverse plan for undo.

Install it in Claude, ChatGPT, Copilot or Gemini

The same skill file works in all four. Pick your tool for the steps.

Install in Claude

  1. 1

    Download the skill

    Fill in the short form to get the .skill file.

  2. 2

    Add it to Claude

    In the Claude app go to Settings → Capabilities → Skills and upload the file. In Claude Code, run the npx command shown after the form, or drop the folder into your project's skills folder.

  3. 3

    Just ask

    Ask in plain language. Claude picks up the skill on its own when the request matches.

How to use it

Once it's installed, ask in plain language. Your AI follows the skill's method when the request matches.

Example prompt 1
Our panel has role stored five different ways. Clean it up and show me what would change.
Example prompt 2
Which attributes in this export are barely filled in, and which ones duplicate each other?
Example prompt 3
Propose fixes for the company size values, but don't change anything until I approve.

What it won't do

It won't write anything to your panel without a person approving each change, and it won't merge values it can't be sure about, such as Employed and Unemployed or two different countries. It won't fill in attributes from transcripts or screener answers until you confirm participant consent covers that use.

Questions

Will it merge values that only look similar?

Not when the difference matters. Values that differ by a negation, a number range, a country or a symbol stay apart, and anything in between goes to a review queue for a person to decide.

Can it change data in my Great Question account?

With the MCP connected it can, but only after you approve each change. It shows the full list first, pauses after the first few writes to check them, and keeps a reverse plan so you can undo.

How does it handle sensitive data?

Attributes in GDPR special categories, such as health or ethnicity, plus any your team lists, need a named reviewer for every merge. It also asks you to confirm consent before using screener answers or transcripts to fill in profiles.

More skills

See all skills