How to use the User Story Writer agent skill (user-stories)

What the user-stories skill makes the agent do

Once the skill is installed, the agent breaks a product feature into user stories. Each story has a short title, a description in the form 'As a [user role], I want to [action], so that [benefit]', a design link placeholder, and 4-6 acceptance criteria written to be testable.

The skill follows the 3 C's (Card, Conversation, Confirmation) and the INVEST criteria (Independent, Negotiable, Valuable, Estimable, Small, Testable). It works from a template and a worked example, and it is told to use plain language.

• Card: a simple title and one-liner
• Conversation: a detailed discussion of intent
• Confirmation: clear acceptance criteria
• Stories are meant to be independent, so they can be developed in any order, and sized for one sprint cycle

How do you install the user-stories skill?

Download the ZIP file from this page. Then choose one of two routes.

For Claude Code, unzip the download into your personal skills folder ~/.claude/skills/, or into a project's .claude/skills/ folder if you want the skill only for that project. The skill's folder contains a file called SKILL.md, which holds all of its instructions.

In the Claude apps, you can instead upload the skill under Settings → Capabilities → Skills.

How do you trigger the skill with a natural request?

You do not need a special command. The skill's description says to use it when writing user stories, breaking down features into backlog items, or defining acceptance criteria, so ask for one of those things directly. Give the agent the feature, the user type and, if you have them, a design link and key assumptions.

The skill's source defines four inputs you can mention: the product or system name, the feature to break down, a link to design files such as Figma or Miro, and key assumptions or context.

• Break this feature into user stories for our checkout: guests can pay without creating an account.
• Write user stories and acceptance criteria for a password reset flow in our mobile app.
• Split this feature into backlog items I can put in next sprint. Design link: [your Figma link].

What does a good output look like?

The skill's own worked example shows the format. A story titled 'Recently Viewed Section' describes an online shopper who wants to revisit items they considered, then lists six acceptance criteria covering when the section appears, which product is excluded, what each card shows and where a click leads.

Read each story against the INVEST list. If a story depends on another one or covers several user types, ask the agent to split or reword it.

• Check each acceptance criterion is observable and testable
• Check the user role is a real group of people you have spoken to
• Check the design link points to actual files, since the skill only inserts a placeholder

What does the skill need, and is it safe to run?

The skill lists no required tools or packages. It does not run scripts and does not use the network, according to the listing. It is a set of written instructions, so the main safety task is reviewing the output rather than managing permissions.

Review the generated stories before they go into a backlog. The output is generated text. It must be checked against real user research and your product decisions, and it is not a verified requirements document.

Tips for better user stories

Give the agent specific context. A feature described with its users, goal and constraints gives more useful stories than a one-line topic. Ask for the stories to be checked against INVEST, and ask for the design link to be filled in when you have one. Keep the plain-language rule: if a story uses jargon the team cannot explain to a newcomer, ask for a rewrite.

• Name the user role precisely, for example 'online shopper' rather than 'user'
• State assumptions up front so they appear in the Conversation part of each story
• Ask for splits when a story is too large for one sprint