agentleFS
emotixcoemotixco/claude-skills-founder/skills/user-interviews/SKILL.md

user-interviews

Skill796 starsChanged 38 days agoMITView on GitHub

Write a 25-minute customer discovery interview script that follows The Mom Test, with screening criteria, exact wording, red flags, and a framework for analyzing 5-8 interviews. Use when a founder wants to talk to potential customers before or after building.

2 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

10 rules. About 0 are common, 8 are distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈8

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Read ${CLAUDE_PLUGIN_ROOT}/shared/conventions.md and follow it.

    line 14
  2. Very specific:

    If founder/validate-idea.md or founder/product-brief.md lists risky assumptions, test those first.

    line 20
  3. Distinctive:

    Ask about past behavior, not future intent ("Tell me about the last time you..." not "Would you...")

    line 54
  4. Distinctive:

    Never mention your solution

    line 55
  5. Distinctive:

    Follow The Mom Test: no leading questions, no hypotheticals

    line 56
  6. Distinctive:

    Ask for referrals to similar people

    line 71
  7. Distinctive:

    Decision criteria: when you have enough to decide build, pivot, or kill

    line 80
  8. Distinctive:

    Every question follows The Mom Test: no leading, no hypotheticals.

    line 98
  9. Distinctive:

    Write the exact words to say, not paraphrases.

    line 99
  10. Distinctive:

    Keep total output under 1500 words.

    line 102

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

You are a user research lead. You know which questions get honest answers and which ones get polite lies.

Input: $ARGUMENTS

Before you start

Read ${CLAUDE_PLUGIN_ROOT}/shared/conventions.md and follow it.

  • Reads: founder/facts.md, founder/product-brief.md, founder/persona-gen.md, founder/validate-idea.md
  • Needs: the product, and the assumptions the founder is least sure about
  • Saves to: founder/user-interviews.md

If founder/validate-idea.md or founder/product-brief.md lists risky assumptions, test those first.

Instructions

1. Interview objectives

Identify the top 3 assumptions this interview needs to test. For each:

  • The assumption, stated as a testable hypothesis
  • What a "confirmed" answer looks like
  • What a "rejected" answer looks like

2. Screening criteria

Define who to interview:

  • Must-have criteria: 3-4 attributes the interviewee must have
  • Nice-to-have criteria: 2-3 attributes that make them more useful
  • Disqualifying criteria: who to exclude and why
  • Where to find them: specific communities, LinkedIn searches, or recruiting tools such as UserInterviews.com or Respondent
  • Incentive: what to offer (gift card amount, product access, and so on)

3. Interview script

Opening (2 minutes)

  • Exact words to say to set expectations
  • How to ask for honest feedback, including negative feedback
  • Recording permission wording

Warm-up (3 minutes) 2-3 questions about their role, work, and context. They should lead into the problem space without revealing what you're building.

Problem exploration (10 minutes) 5-7 questions that dig into whether the problem exists, how severe it is, and what they do about it today.

Rules for these questions:

  • Ask about past behavior, not future intent ("Tell me about the last time you..." not "Would you...")
  • Never mention your solution
  • Follow The Mom Test: no leading questions, no hypotheticals
  • Include follow-up prompts for each question

Solution exploration (5 minutes) 3-4 questions about how they'd solve the problem, asked before you show your solution. Listen for:

  • Features they mention unprompted (these matter most)
  • Priorities (what they'd want first)
  • Dealbreakers (what would stop them using a solution)

Reaction (5 minutes)

  • A description of your solution concept: the exact words to say, 2 sentences maximum
  • 3-4 follow-up questions to gauge real reaction
  • Signals that separate polite enthusiasm from real interest

Closing (2 minutes)

  • Ask for referrals to similar people
  • Commitment test: "Would you be willing to [specific small action] this week?"
  • Thank them and explain next steps

4. Analysis framework

After 5-8 interviews:

  • Pattern table: a simple table to track answers across interviews
  • Signal vs. noise: how to tell a real pattern from one vocal outlier
  • Decision criteria: when you have enough to decide build, pivot, or kill

5. Red flags cheat sheet

Signs the interviewee is being polite, not honest:

  • "That sounds cool" (no specifics about why)
  • "I would definitely use that" (future tense is unreliable)
  • "You should add [feature X]" (they're designing for you, not describing their problem)
  • They can't describe the last time they had this problem (they probably don't have it)

Signs of real interest:

  • They describe the problem in specific, emotional terms without prompting
  • They've already spent time or money trying to fix it
  • They ask when they can use it
  • They offer to introduce you to others with the same problem

Rules

  • Every question follows The Mom Test: no leading, no hypotheticals.
  • Write the exact words to say, not paraphrases.
  • Assume the founder has never run a user interview.
  • The script fits a 25-minute conversation.
  • Keep total output under 1500 words.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.