Claude Code Plugins

Community-maintained marketplace

Feedback

This skill should be used when the user asks to "create tasks", "break down story into tasks", "define tasks", "what tasks are needed", "write acceptance criteria", "implementation tasks", "task list", "create work items", "technical tasks", "work breakdown", "decompose story", "story to tasks", or when decomposing user stories into specific, executable tasks with clear acceptance criteria for GitHub Projects.

Install Skill

1Download skill
2Enable skills in Claude

Open claude.ai/settings/capabilities and find the "Skills" section

3Upload to Claude

Click "Upload skill" and select the downloaded ZIP file

Note: Please verify skill by going through its instructions before using it.

SKILL.md

name task-breakdown
description This skill should be used when the user asks to "create tasks", "break down story into tasks", "define tasks", "what tasks are needed", "write acceptance criteria", "implementation tasks", "task list", "create work items", "technical tasks", "work breakdown", "decompose story", "story to tasks", or when decomposing user stories into specific, executable tasks with clear acceptance criteria for GitHub Projects.

Task Breakdown

Quick Actions & Routing

User Intent Action Resource
Reviewing story first Understand user goal and value Step 1: Review the User Story
Identifying tasks Map implementation layers references/task-layers.md
Applying task patterns Use CRUD/feature/integration patterns references/task-patterns.md
Defining acceptance criteria Use testable statement format Step 4: Define Acceptance Criteria
Creating task issues Use template references/task-template.md
Viewing examples Load sample task set examples/example-task-set.md

Command Integration

The /re:create-tasks command guides task creation in GitHub Projects. This skill provides the methodology for breaking down stories into tasks—including task patterns, acceptance criteria templates, and layer-by-layer decomposition techniques. Load this skill for deeper understanding of task breakdown concepts or when you need guidance beyond what the command provides.

Overview

Task breakdown transforms user stories into concrete, executable work items that can be assigned, tracked, and completed. Tasks represent the actual implementation steps needed to deliver a user story, each with clear acceptance criteria. This skill guides the process of decomposing stories into well-defined tasks suitable for GitHub issue tracking.

Purpose

Tasks are the execution layer in the requirements hierarchy:

  • Above: User Stories (user-facing functionality)
  • Tasks: Specific implementation steps
  • Below: (Nothing—tasks are the lowest level)

Well-defined tasks:

  • Represent discrete units of work (2-8 hours typical, up to 1-2 days max)
  • Have clear, testable acceptance criteria
  • Can be assigned to individuals
  • Track progress toward story completion
  • Enable accurate status reporting

Prerequisite

User story must exist before creating tasks. If no story exists, use the user-story-creation skill first.

Task Characteristics

What Makes a Good Task

Specific and Concrete:

  • Clear, unambiguous description
  • Obvious what needs to be done
  • No interpretation needed

Right-sized:

  • 2-8 hours of work (up to 1-2 days maximum)
  • Small enough to complete in a single sitting when possible
  • Large enough to deliver meaningful progress

Testable:

  • Has clear acceptance criteria
  • Can verify when complete
  • Observable outcome

Assignable:

  • One person can own and complete it
  • Doesn't require coordinating multiple people

Valuable:

  • Contributes toward completing the story
  • Represents real progress

Task Breakdown Process

Step 1: Review the User Story

Understand the story being implemented:

Key Actions:

  • Read story issue in GitHub Projects
  • Understand the user goal and value
  • Review acceptance criteria
  • Note any constraints or assumptions

Step 2: Identify Implementation Layers

Break down work by typical software layers. See references/task-layers.md for detailed guidance, examples, and checklists for each layer.

Common Layers:

Layer Focus
Frontend/UI Components, styling, interactions, state
Backend/API Endpoints, business logic, validation
Data/Database Schema, migrations, queries, indexes
Testing Unit, integration, E2E, manual
Documentation API docs, user guides, code comments
DevOps Configuration, deployment, monitoring
Research/Spike Evaluation, prototypes, feasibility

Step 3: Apply Common Task Patterns

Use proven patterns as starting points for task breakdown. See references/task-patterns.md for detailed guidance, examples, and checklists.

Common Patterns:

Pattern When to Use
CRUD Operations Building entity management (create, read, update, delete)
Feature Implementation Adding new user-facing capabilities
Integration Connecting to external services or APIs
Refactoring Improving code structure without changing behavior
Bug Fix Fixing production issues
Migration Moving to new systems or technologies

Step 4: Define Acceptance Criteria

For each task, specify clear success conditions:

Format:

Use specific, testable statements:

  • "Component renders correctly with props X, Y, Z"
  • "API endpoint returns 200 status with correct data structure"
  • "Database migration runs without errors and creates table T"
  • "Unit tests achieve >80% code coverage for module M"

Key Elements:

  • What should exist/work when task is complete
  • How to verify it works
  • Any performance or quality standards

Step 5: Sequence and Dependencies

Order tasks logically:

Dependency Analysis:

  • Which tasks must come first?
  • What's the critical path?
  • Can any tasks be done in parallel?

Typical Sequence:

  1. Data/schema changes (if needed)
  2. Backend implementation
  3. Frontend implementation
  4. Integration and testing
  5. Documentation

Mark Dependencies:

  • Note which tasks block others
  • Identify tasks that can start immediately
  • Plan for parallel work when possible

Step 6: Create Task Issues in GitHub Projects

For each task, create a GitHub issue in the relevant GitHub Project:

Issue Title: "[Clear, action-oriented task description]"

Issue Description: Task details and acceptance criteria using template

Custom Fields:

  • Type: Task
  • Priority: (inherited from story)
  • Status: Not Started

Labels:

  • type:task
  • Technical labels (frontend, backend, database, testing, docs)

Parent: Link to Story issue as parent

Estimate: (optional) Hour or story point estimate

Task Templates and Examples

For detailed task templates and examples by type (Frontend, Backend, Database, Testing, Documentation, Research/Spike), see references/task-template.md.

Best Practices

Keep Tasks Focused

One clear objective per task:

❌ "Implement campaign filtering and sorting and export" ✅ "Implement campaign date filtering" ✅ "Implement campaign sorting by name" ✅ "Implement campaign CSV export"

Use Action-Oriented Titles

Start with verbs:

Good:

  • "Create campaign filter component"
  • "Implement date validation logic"
  • "Add database index for performance"
  • "Write unit tests for filter module"

Poor:

  • "Campaign filtering" (vague)
  • "Database" (what about it?)
  • "Tests" (what kind? for what?)

Include Technical Context

Help the implementer:

  • Reference existing code patterns to follow
  • Note relevant files or modules
  • Mention design system components to use
  • Link to related tasks or documentation

Balance Granularity

Not too big, not too small:

Too big: "Implement entire campaign management system" (this is a story or epic)

Too small: "Import React" (this is a substep, not a task)

Just right: "Create CampaignList component with sorting and filtering props"

Always Include Acceptance Criteria

Never create a task without clear success conditions:

  • Minimum 3-5 criteria per task
  • Make them specific and testable
  • Include both functional and quality criteria

Common Pitfalls to Avoid

Tasks Without Acceptance Criteria

Every task needs testable success conditions:

❌ "Work on campaign filtering" ✅ "Implement date filter UI" + 5 specific acceptance criteria

Tasks Too Large

Watch for multi-day or multi-person tasks:

  • Split tasks larger than 2 days
  • Each task should be completable by one person
  • If coordination needed, it's probably too big

Mixing Concerns

One task, one focus area:

❌ "Implement filter UI and backend API and database schema" ✅ Three separate tasks for UI, API, and database

Vague Descriptions

Be specific about what needs to be done:

❌ "Fix bugs" ✅ "Fix date validation bug where end date before start date is allowed"

Missing Dependencies

Note what must be done first:

  • Database schema before backend code
  • Backend API before frontend integration
  • Core functionality before tests (usually)

Quick Reference: Task Breakdown Flow

  1. Review Story → Understand user goal, value, acceptance criteria
  2. Identify Layers → Frontend, backend, database, testing, docs, infrastructure
  3. Apply Patterns → Use common task patterns as starting points
  4. Define Acceptance Criteria → Specify testable success conditions for each task
  5. Sequence → Order tasks, note dependencies
  6. Create Issues → Add to GitHub Projects as children of story
  7. Assign → (Optional) Assign tasks to team members
  8. Execute → Begin work, update task status as progress is made

Integration with GitHub Issues (GitHub Projects)

Task Issue Format

Title: Clear, action-oriented description

Labels:

  • type:task
  • Technical area (frontend, backend, database, etc.)
  • Priority (inherited from story)

Assignment: Person responsible

Estimate: Optional hours or points

Parent Link: User story issue

Acceptance Criteria: In issue description

Tracking Progress

Tasks enable granular progress tracking:

  • Story progress = % of tasks complete
  • Epic progress = % of stories complete (via tasks)
  • Vision progress = % of epics complete (via stories → tasks)

Full traceability: Vision → Epic → Story → Task

Reference Files

Load references as needed:

Reference When to Load Path
task-layers.md Breaking down stories by implementation layer references/task-layers.md
task-patterns.md Applying common task patterns (CRUD, feature, integration) references/task-patterns.md
task-template.md Creating task issue content or defining acceptance criteria references/task-template.md

Examples

Working examples that can be copied and adapted:

Example Use Case Path
example-task-issue.md Creating a single task with full detail examples/example-task-issue.md
example-task-set.md Viewing related tasks for a story examples/example-task-set.md

Related Skills

Load these skills when task work reveals needs beyond this skill's scope:

Task Context Load Skill Routing Trigger
No stories exist or story is unclear user-story-creation User needs to create or refine user stories
Task priorities need to be established prioritization User needs to apply MoSCoW framework to tasks
Tasks reveal implementation issues requirements-feedback User needs to gather feedback or refine requirements