Guide · Agent skill
How to use the User Persona Builder agent skill
Updated · written with AI from the project's own documentation and checked against our catalogue data
The User Persona Builder is an agent skill from the phuryn/pm-skills collection, named user-personas. Once installed, it tells an AI agent to build three distinct user personas from research you supply, such as survey responses, interview transcripts or CSV and Excel files. This guide explains what the skill does, how to install it, how to trigger it and what to watch for, so you can get usable persona documents from your own data.
User Persona Builderphuryn/pm-skills · MIT Download the skill Quick start
- 1Download the skill. Get the user-personas ZIP from our page. It contains SKILL.md.
- 2Unzip into a skills folder. Place the folder in ~/.claude/skills/ for personal use or in the project's .claude/skills/ folder.
- 3Or upload it in the Claude apps. Go to Settings → Capabilities → Skills and upload the ZIP.
- 4Prepare your research. Collect survey exports as CSV or Excel, and save interview or notes as text files the agent can read.
- 5Ask for personas. Make a natural request that names your product and mentions building personas from the research.
- 6Review and verify. Check the personas against the source data, follow up on flagged gaps and confirm quotes.
What is an agent skill, and what does this one make the agent do?
An agent skill is a folder containing a SKILL.md file with instructions that an AI agent loads when a task matches the skill's description. This one lives at pm-market-research/skills/user-personas in the phuryn/pm-skills repository and is licensed MIT.
The skill asks the agent to read your research data, look for recurring patterns, group similar users into segments and write three personas. For each persona it covers the primary job-to-be-done, the top three pain points, the top three desired gains, one unexpected behavioural insight, and how your product fits that persona's needs. The instructions also ask the agent to ground every claim in the data and to flag gaps where more research is needed.
Install the user-personas skill
Download the ZIP of the skill from our page. The package contains a single file, SKILL.md. Unzip it so the skill folder sits in one of these locations:
- Personal skills folder: ~/.claude/skills/
- Project-specific: the .claude/skills/ folder inside your project
- Or upload the ZIP in the Claude apps under Settings → Capabilities → Skills
~/.claude/skills/Trigger the skill with a natural request
You do not need a special command. The skill's description says to use it when building personas from survey data, creating user profiles from research, or segmenting users for product decisions. Ask for personas in those terms and name your product or segment, because the instructions insert that name into the analysis.
Provide the research in a form the agent can read: CSV, Excel, or text research notes such as survey responses or interview transcripts.
- "Create three user personas for our invoicing app from this survey export.csv"
- "Build personas from these five interview transcripts for the onboarding redesign."
- "Segment our users into personas for product decisions using the notes in research/."
What you get back
Each persona follows the same structure: a name and demographics (age range, role, company size for B2B, key characteristics), the primary job-to-be-done with context and frequency, pain points with impact and severity, desired gains with how success is measured, one unexpected insight with why it matters, and a product fit assessment that lists friction points and unmet needs.
The instructions ask for direct quotes from the research where they exist and for personas that are distinct and non-overlapping where possible. Treat the output as a draft to check against your data, not a finished deliverable.
Safety notes: scripts and network
This skill does not run scripts and does not use the network, according to its listing. It works by having the agent read the files you give it. That means the data you upload is handled by your AI agent, so check that your agent setup is allowed to process that material, especially for customer personal data.
- Runs scripts: no
- Uses network: no
- Remove sensitive identifiers from transcripts and survey exports before sharing them with the agent where possible
Tips for better personas
Output quality depends on input quality. Thin or biased research produces thin or biased personas. The skill is designed to flag gaps, so read those flags as a to-do list for further research.
Use a mix of sources if you have them. Survey data shows breadth, while interview transcripts supply the quotes and reasoning behind the numbers.
- Give the agent real data only; the skill is built to avoid invented personas
- Ask for the gaps section explicitly if you want the research follow-ups listed
- Check each persona's quotes against the original files before using them in a deck
What people use it for
- Turning a customer survey export into three persona documents for a product review
- Summarising a set of user interview transcripts into jobs-to-be-done and pain points
- Segmenting a small startup's raw feedback into distinct user groups before choosing a roadmap focus
- Creating a consistent persona format across several research studies run by one team
Frequently asked questions
How do I install the user-personas skill?
Download the ZIP from our page and unzip it into ~/.claude/skills/ for personal use, or into a project's .claude/skills/ folder. You can also upload the ZIP in the Claude apps under Settings → Capabilities → Skills.
Is the user personas skill free?
The skill is licensed under MIT, as stated in its listing. You still need an AI agent that can read your files to use it.
What do I need to run the user personas skill?
You need an AI agent that can read the files you provide, such as CSV, Excel or text research notes. The skill does not run scripts and does not use the network.
Does the skill work with survey data and interview transcripts?
Yes. The skill is designed for survey data, interview transcripts and other research files. It asks the agent to extract patterns, demographics, motivations and behaviours from them.
Can it create personas without any research data?
It is not meant for that. The skill grounds personas in evidence, and invented personas would be misleading, so provide real research data first.
How is this different from writing personas by hand?
The skill applies a consistent structure to each persona and asks the agent to ground claims in your data and flag gaps. Your own judgement is still needed to check the output against the source material.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
This skill suits product managers, founders and researchers who already have survey or interview data and want consistent, data-grounded personas quickly. It is a poor fit if you have no real research, since the skill's output depends entirely on the quality of what you provide and should not replace primary research.
Download the skill