Requirements Skill | Agent Skills

Requirements

Transform clarified intent into actionable user stories with acceptance criteria.

Input

Expect intent output containing:

Problem statement
Users and their needs
Constraints (technical, business, timeline)
Success criteria
Coding standards (if gathered)

Process

1. Identify story candidates

Review the intent and identify discrete pieces of functionality:

Each user need typically maps to one or more stories
Each success criterion should be achievable by one or more stories
Break large features into smaller, independently deliverable stories

2. Write user stories

For each story, use the standard format:

As a [user persona],
I want [goal/desire],
So that [benefit/value].

Guidelines:

Stories should be independent where possible
Stories should be negotiable (not contracts)
Stories should be valuable to the user
Stories should be estimable
Stories should be small enough to complete in one iteration
Stories should be testable

3. Define acceptance criteria

For each story, write acceptance criteria using Given/When/Then:

Given [initial context/state],
When [action is taken],
Then [expected outcome].

Guidelines:

Cover the happy path first
Include key edge cases
Be specific and testable
Avoid implementation details
Each criterion should be independently verifiable

4. Identify dependencies

Mark dependencies between stories:

Technical dependencies (A must exist before B can work)
Logical dependencies (A provides context for B)
Avoid circular dependencies

5. Prioritise

Assign priority based on:

User value (higher = more valuable)
Risk reduction (higher = reduces more uncertainty)
Dependencies (blocked stories rank lower)
Quick wins (small high-value items rank higher)

Use priority numbers: 1 = highest, incrementing for lower priority.

6. Estimate complexity

Assign t-shirt sizes:

S - Few hours, single component
M - A day or two, few components
L - Several days, multiple components
XL - Week or more, significant complexity (consider splitting)

Output

Generate backlog in this format:

backlog:
  - id: "US-001"
    title: "Short descriptive title"
    description: |
      As a [user],
      I want [goal],
      So that [benefit].
    acceptance_criteria:
      - "Given [context], when [action], then [result]"
      - "Given [context], when [action], then [result]"
    priority: 1
    estimate: "M"
    dependencies: []

  - id: "US-002"
    title: "Another story"
    description: |
      As a [user],
      I want [goal],
      So that [benefit].
    acceptance_criteria:
      - "Given [context], when [action], then [result]"
    priority: 2
    estimate: "S"
    dependencies: ["US-001"]

Story ID format

Use US-NNN format with zero-padded numbers starting from 001.

Validation

Before finalising, verify:

[ ] Every success criterion from intent is addressed by at least one story
[ ] Every user persona has at least one story addressing their needs
[ ] No orphan stories (stories that don't trace back to intent)
[ ] Dependencies form a directed acyclic graph (no cycles)
[ ] Acceptance criteria are testable without knowing implementation
[ ] XL stories have been considered for splitting

Presentation

Present the backlog to the user:

Summarise: "Based on the intent, I've identified N stories"
Show dependency graph if complex
List stories in priority order with estimates
Ask for confirmation before saving to manifest

Manifest update

Once approved, update .sdlc/manifest.yaml:

manifest:
  backlog:
    - id: "US-001"
      title: "..."
      # ... full story details
  current_story: null  # Ready for story selection

Tips

When in doubt, make stories smaller rather than larger
Acceptance criteria should read like a test specification
If a story has more than 5-7 acceptance criteria, consider splitting
Technical tasks (refactoring, infrastructure) can be stories if they deliver user value indirectly
Include "spike" stories for research when uncertainty is high

Agent Skills: Requirements

Install this agent skill to your local

Skill Files