Acceptance Criteria Definer

Overview

Define clear, testable acceptance criteria that ensure shared understanding between product, development, and QA. Criteria should be specific enough to test but flexible enough to allow implementation decisions.

Core principle: If you can't write a test for it, it's not a valid acceptance criterion.

When to Use

• Completing draft user stories before refinement
• Clarifying vague requirements with stakeholders
• Preparing stories for sprint planning
• Defining done for complex features

Output Format

acceptance_criteria:
  story_reference: "[Story ID]"
  story_title: "[Story title]"
  
  preconditions:
    - "[System state required before testing]"
  
  scenarios:
    happy_path:
      - id: "AC-01"
        title: "[Scenario name]"
        given: "[Initial context]"
        when: "[Action performed]"
        then: "[Expected outcome]"
        priority: "[Must | Should | Could]"
    
    alternate_paths:
      - id: "AC-02"
        title: "[Alternate scenario]"
        given: "[Different context]"
        when: "[Same or different action]"
        then: "[Different outcome]"
        priority: "[Must | Should | Could]"
    
    error_handling:
      - id: "AC-03"
        title: "[Error scenario]"
        given: "[Context that causes error]"
        when: "[Action attempted]"
        then: "[Error handling behavior]"
        priority: "[Must | Should | Could]"
    
    edge_cases:
      - id: "AC-04"
        title: "[Boundary or edge case]"
        given: "[Edge condition]"
        when: "[Action at boundary]"
        then: "[Expected behavior]"
        priority: "[Should | Could]"
  
  non_functional:
    performance:
      - "[Response time requirements]"
    security:
      - "[Security constraints]"
    accessibility:
      - "[A11y requirements]"
  
  out_of_scope:
    - "[Explicitly not covered]"
  
  validation_approach:
    automated: "[What can be automated]"
    manual: "[What requires manual testing]"
    
  coverage_assessment:
    happy_path: "[Complete | Partial | Missing]"
    error_handling: "[Complete | Partial | Missing]"
    edge_cases: "[Complete | Partial | Missing]"
    non_functional: "[Complete | Partial | Missing]"

Given-When-Then Guidelines

Given (Preconditions)

• State the starting context
• Include relevant data state
• Specify user authentication/authorization if relevant
• Be specific about system state

Good: "Given a logged-in user with admin privileges and an existing product with ID 123" Bad: "Given the system is ready"

When (Actions)

• Single action per criterion
• Use active voice
• Specify exact interaction
• Include relevant parameters

Good: "When the user clicks 'Delete' and confirms the deletion dialog" Bad: "When the user does the delete"

Then (Outcomes)

• Observable result
• Specific and measurable
• Include relevant state changes
• Consider user feedback and system state

Good: "Then the product is removed from the list, a success message displays for 3 seconds, and the product count decreases by 1" Bad: "Then it works"

Coverage Categories

Happy Path (Required)

The expected, successful flow:

- id: "AC-01"
  title: "Successful order submission"
  given: "User has items in cart and valid payment method"
  when: "User clicks 'Place Order'"
  then: "Order is created, confirmation email sent, inventory updated"

Alternate Paths (Required)

Valid but non-primary flows:

- id: "AC-02"
  title: "Order with promo code"
  given: "User has items in cart and valid promo code"
  when: "User applies promo code and places order"
  then: "Discount applied, order total updated, order created"

Error Handling (Required)

How system handles failures:

- id: "AC-03"
  title: "Payment declined"
  given: "User has items in cart and payment method that will be declined"
  when: "User attempts to place order"
  then: "Error message displayed, order not created, user can retry with different payment"

Edge Cases (Recommended)

Boundary conditions:

- id: "AC-04"
  title: "Cart at maximum item limit"
  given: "User has 99 items in cart (max is 100)"
  when: "User adds one more item"
  then: "Item added successfully, cart shows 100 items"

- id: "AC-05"
  title: "Cart exceeds maximum"
  given: "User has 100 items in cart"
  when: "User attempts to add another item"
  then: "Error message explains limit, item not added"

Priority Levels

Common Patterns

Form Validation

scenarios:
  happy_path:
    - id: "AC-01"
      title: "Valid form submission"
      given: "User on registration form"
      when: "User enters valid email, password meeting requirements, and submits"
      then: "Account created, confirmation email sent, user redirected to dashboard"
  
  error_handling:
    - id: "AC-02"
      title: "Invalid email format"
      given: "User on registration form"
      when: "User enters 'notanemail' in email field and attempts submit"
      then: "Inline error appears: 'Please enter a valid email address', form not submitted"
    
    - id: "AC-03"
      title: "Password too weak"
      given: "User on registration form"
      when: "User enters password shorter than 8 characters"
      then: "Inline error appears with password requirements, form not submitted"

Search/Filter

scenarios:
  happy_path:
    - id: "AC-01"
      title: "Search returns results"
      given: "User on product listing with 100 products"
      when: "User searches for 'laptop'"
      then: "Results filtered to products containing 'laptop', result count displayed"
  
  edge_cases:
    - id: "AC-02"
      title: "No search results"
      given: "User on product listing"
      when: "User searches for 'xyznonexistent'"
      then: "Empty state displayed: 'No products found', suggestion to modify search"
    
    - id: "AC-03"
      title: "Empty search"
      given: "User on product listing with active search filter"
      when: "User clears search field"
      then: "All products displayed, filter cleared"

Anti-Patterns

Completeness Checklist

Before finalizing:

• Happy path defined with at least one scenario
• Alternate valid paths identified
• Error handling for likely failures
• Edge cases at boundaries
• Non-functional requirements if applicable
• All scenarios are testable
• Priority assigned to each scenario
• Out of scope explicitly stated