Guide · Agent skill
How to use the Summarize Customer Interview agent skill
Updated · written with AI from the project's own documentation and checked against our catalogue data
The summarize-interview skill is an agent skill that turns a raw customer interview transcript into a short, structured summary. It is built for product managers, founders and researchers who run discovery interviews and want consistent notes. After this guide you will know what the skill does, how to install it, how to ask for it in plain language, and what to check before trusting the output.
Quick start
- 1Download the skill ZIP. Download the summarize-interview ZIP from this page.
- 2Unzip into a skills folder. Unzip it into ~/.claude/skills/ for personal use or .claude/skills/ inside a project.
- 3Or upload in the Claude apps. Go to Settings, then Capabilities, then Skills, and upload the skill.
- 4Attach or paste the transcript. Provide the interview text or an attached file, and name the product or study.
- 5Ask for a summary in plain language. Request an interview summary or processing of discovery interviews.
- 6Review the markdown output. Check the saved summary, especially quotes, names and any "-" entries.
What an agent skill is and what this one makes the agent do
An agent skill is a set of instructions packaged as a folder with a SKILL.md file. When it is installed, the agent can load these instructions and follow them for matching tasks. This skill, from the phuryn/pm-skills repository (pm-product-discovery/skills/summarize-interview, MIT licence), tells the agent to summarize a customer interview using a fixed template.
The agent reads the full transcript, whether it is pasted or attached as text, PDF or an audio transcription, and fills in the template. The summary covers the customer's background, their current solution, the jobs they are trying to get done, problems and likes with that solution, key insights or quotes, and action items with a date, owner and action. The result is saved as a markdown document in your workspace.
Install the summarize-interview skill
You can install the skill in one of two ways. Download the ZIP of the skill from this page and unzip it into one of these locations: ~/.claude/skills/ for personal use across projects, or .claude/skills/ inside a specific project folder if you want it available only there.
Alternatively, in the Claude apps, go to Settings, then Capabilities, then Skills, and upload the skill there.
- Personal install: unzip into ~/.claude/skills/
- Project install: unzip into .claude/skills/ in your project
- Claude apps: Settings → Capabilities → Skills, then upload
Trigger the skill with natural requests
You do not need a special command. The skill's description says it applies when you are processing interview recordings or transcripts, synthesizing discovery interviews, or creating interview summaries, so ask for that work in ordinary language. Attach the transcript file or paste the text, and name the product or study so the context is clear, since the skill uses the topic you give it.
- Summarize this customer interview transcript for our onboarding discovery.
- Create an interview summary from the attached file for the pricing research.
- Process these discovery interviews and list the action items with owners.
Example: what the output looks like
The skill follows a fixed template, so every summary has the same headings. Missing information is shown as "-" rather than guessed. Action items use the pattern date, owner, action. The example below is the format from the skill itself.
- Date and time of the interview
- Participants with full names and roles
- Current solution, likes and problems with job, desired outcome, importance and satisfaction
- Key insights and notable quotes
- Action items as date, owner, action
**Action Items**:
- [Date, Owner, Action — e.g., "2025-01-15, Paweł Huryn, Follow up with customer about pricing"]What the skill needs and safety notes
The skill lists no required tools or dependencies. It does not run scripts and does not use the network, according to its assessment. The skill sends the transcript content to the model you are using, so treat it like any text you share with that model.
The skill does not describe any redaction of personal data. For legal, medical or compliance-sensitive recordings, remove identifying details yourself before sharing, or choose a different process.
- No listed requirements
- Runs no scripts
- Uses no network access by the skill itself
- Transcript content is sent to the model; no redaction is described
Tips for better summaries
The summary can only reflect what is in the transcript. Poor transcription or an incomplete interview leads to more "-" entries, so clean up obvious errors first if you can. Review quotes against the original, because the skill asks for notable quotes and you should confirm they are accurate before sharing.
When you process a batch, summarize one interview at a time so each markdown file stays tied to one customer. Then compare the action items across files yourself.
- Start with a clean, complete transcript
- Check quotes and names against the source
- Summarize one interview per request for clean files
What people use it for
- Synthesizing a batch of discovery interviews for a product team
- Capturing follow-up actions with owners right after a customer call
- Keeping consistent interview notes across founders or researchers
- Pulling jobs to be done and satisfaction signals from several customers
Frequently asked questions
How do I install the summarize-interview skill?
Download the skill ZIP from this page and unzip it into ~/.claude/skills/ for personal use or into .claude/skills/ in a project. In the Claude apps you can instead go to Settings, then Capabilities, then Skills, and upload it.
Is the summarize-interview skill free?
The skill is published under the MIT licence in the phuryn/pm-skills repository. This page does not list any price, so check the repository for licence terms and any costs of the agent you use.
What does the skill need to run?
The skill lists no required dependencies. It does not run scripts and does not use the network, though the transcript content is sent to the model you are using.
Can I use it on sensitive or legal interview recordings?
The skill's assessment advises caution for legal, medical or compliance-sensitive recordings, because it sends transcript content to the model and describes no redaction of personal data. Remove identifying details yourself first or use another process.
Why does my summary contain dashes instead of answers?
The template uses "-" when information is unavailable. This usually means the transcript is incomplete or poorly transcribed, since the summary only reflects what is in the text.
How does this compare with writing interview notes by hand?
The skill gives a consistent structure with jobs to be done, satisfaction, insights and action items each time. It does not replace reviewing the transcript yourself, so check quotes and names against the source.
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
Use this skill if you run regular discovery interviews and want the same structured summary and follow-up actions every time, with no scripts or network access needed. Pick something else if your transcripts are poor quality or your recordings are legal, medical or compliance-sensitive, since the skill offers no redaction of personal data.
Download the skill