Moderator scorecard

A Claude skill that shows you, from your own interview transcript, how much you talked, interrupted or led the participant, so you can get better at moderating.

Research opsOctober 12, 2026

Download the skill

Enter your details to download the skill file.

What it does

Interviewers rarely find out how much they talk, interrupt or lead the participant.

You give it a transcript of an interview you ran. It tells you how much of the session you talked compared with the participant, your longest speeches, any interruptions, and questions that pushed toward an answer. Use it to review your own sessions and improve. It isn't for managers grading people.

What you get back

  • Talk-time share for each side, by duration where timings allow and by word count otherwise.
  • Your longest monologues, quoted, so you can see what they were.
  • Interruptions counted only where the speech really overlaps.
  • Leading, presuming and tag questions, quoted with their timestamps.
  • Guide coverage as information, so a productive tangent isn't marked as a problem.
  • A plain note on what couldn't be measured from this transcript.

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
Score my moderation on yesterday's interview. Where did I talk too much?
Example prompt 2
Check this session transcript for leading questions.
Example prompt 3
This was a think-aloud usability session. What should I look at in my moderation?

What it won't do

It isn't for performance evaluation. It measures one session for the person who ran it. It also depends on the speaker labels: diarisation errors can flip a talk-time result, so check who is who before reading any number. And it can't hear tone, so a question that reads as leading may have landed differently in the room.

Questions

What transcript format does it need?

WebVTT is best, because each cue has a start and an end time, which pauses and interruptions need. A plain timestamped transcript works for everything else, including Otter-style exports.

How does it know who the moderator is?

It looks for a speaker labelled as moderator, interviewer, researcher or facilitator, or one who asks clearly more questions than anyone else. If that's ambiguous it stops and asks you to name the moderator, and it supports two moderators.

Should my team use it to review each other?

Only if the person being reviewed asked for it. It is meant for self-review and coaching. Check that your participant consent covers quality review of transcripts before you run it.

More skills

See all skills