mcpField
A
audit-stories

audit-stories

github.com/owl-listenerTesting & quality10873 installsUpdated 20w ago
Repository

About audit-stories

A skill for reviewing existing user stories for accessibility gaps and adding concrete acceptance criteria to those stories instead of creating separate accessibility tickets. It covers varied input, sensory, and cognitive needs and produces a story-by-story audit report.

Key features

  1. 01Check keyboard and screen reader support
  2. 02Review visual, motor, and cognitive access
  3. 03Identify assumptions about abilities and context
  4. 04Add criteria to existing stories
  5. 05Generate summary and story-level findings

Use cases

  1. 01Audit a sprint backlog of user stories
  2. 02Review accessibility across a feature's story set
  3. 03Write ready-to-use criteria for stories with gaps
  4. 04Find recurring design omissions across stories

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/owl-listener/inclusive-design-skills --skill audit-stories

After installing, type/audit-storiesin Claude Code to confirm it loaded.

Details

From the project README

Audit User Stories for Inclusion

Review an existing set of user stories and identify where disability and diverse abilities have been overlooked — then add the missing criteria without creating separate accessibility tickets.

Process

Step 1: Gather Stories

Collect the user stories to audit. This could be a sprint backlog, a feature's full story set, or an epic.

Step 2: Check Each Story

Using inclusive-user-stories, review each story's acceptance criteria for:

  • Keyboard operability
  • Screen reader support
  • Visual accessibility (contrast, colour independence, zoom)
  • Motor accessibility (target size, timing, precision)
  • Cognitive accessibility (plain language, error recovery, undo)

Flag any story where one or more of these are missing.

Step 3: Check for Assumption Bias

Using edge-case-identification, for each story ask:

  • What input method does this assume?
  • What sensory ability does this assume?
  • What cognitive capacity does this assume?
  • What context does this assume?
  • What "edge cases" have been dismissed that are daily reality for some users?
Step 4: Add Missing Criteria

For each flagged story, write the specific acceptance criteria that need to be added. Don't create separate stories — add directly to the existing ones.

Output

Deliver an audit report:

  1. Summary — number of stories reviewed, number with gaps, most common gap type
  2. Story-by-story findings — each story with its missing criteria listed
  3. Added criteria — the specific acceptance criteria to add to each story, ready to paste into the backlog
  4. Patterns — recurring gaps that suggest a systemic issue (e.g., "no story in this sprint mentions keyboard access")

Comments

Markdown supported