Claude Code Plugins

Community-maintained marketplace

Feedback

Use during Phase 4 of spec creation to break specification into organized task groups with dependencies - analyzes requirements to determine natural groupings, presents structure for validation, and creates detailed actionable tasks

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 tasks-planning
description Use during Phase 4 of spec creation to break specification into organized task groups with dependencies - analyzes requirements to determine natural groupings, presents structure for validation, and creates detailed actionable tasks

Tasks Planning

What It Does

  1. Analyzes spec and requirements
  2. Determines natural task groupings by layer/skill
  3. Presents structure for validation
  4. Creates detailed tasks.md with dependencies
  5. Notes reusability and visual references

Tasks are specific, verifiable, and ordered by dependencies.

The Process

Step 1: Load Specification

SPEC="[provided by workflow]"

cat "$SPEC/spec.md"
cat "$SPEC/planning/requirements.md"
ls -la "$SPEC/planning/visuals/"

Analyze:

  • Requirements and dependencies
  • Technical layers (database, API, frontend)
  • Existing code to leverage
  • Visual designs to implement

Step 2: Determine Task Groups

Identify natural groupings:

Common patterns:

  • Database Layer (models, migrations)
  • API Layer (controllers, routes, auth)
  • Frontend Components (UI, forms, pages)
  • Integration & Testing (end-to-end verification)

Consider dependencies:

  • Database before API (API needs models)
  • API before Frontend (Frontend needs endpoints)
  • Implementation before Testing

Present structure:

Based on spec, I'm organizing tasks into:

1. **Database Layer**
   - Models, migrations, associations
   - Dependencies: None (foundation)

2. **API Layer**
   - Controllers, endpoints, auth
   - Dependencies: Database Layer

3. **Frontend Components**
   - UI, forms, pages
   - Dependencies: API Layer

4. **Integration & Testing**
   - End-to-end verification
   - Dependencies: All implementation

This follows dependency order: foundation → API → interface → verification

Does this grouping make sense, or organize differently?

WAIT for confirmation.

Step 3: Present Task Breakdown

After grouping approved:

Detailed task breakdown:

## Task Group 1: Database Layer
**Dependencies:** None

- [ ] 1.0 Complete database layer
  - [ ] 1.1 Create [Model] with validations
    - Fields: [list]
    - Validations: [list]
    - Reuse pattern from: [existing if applicable]
  - [ ] 1.2 Create migration for [table]
    - Add indexes for: [fields]
    - Foreign keys: [relationships]
  - [ ] 1.3 Set up associations
    - [Model] has_many [related]
    - [Model] belongs_to [parent]

**Acceptance Criteria:**
- Models pass validation tests
- Migrations run successfully
- Associations work correctly

---

## Task Group 2: API Layer
**Dependencies:** Task Group 1

[Similar structure]

---

[Continue for all groups]

Does this breakdown capture all requirements?

WAIT for confirmation.

Step 4: Create tasks.md

After breakdown approved:

cat > "$SPEC/tasks.md" <<'EOF'
# Task Breakdown: [Feature Name]

## Overview
Total Task Groups: [count]
Total Tasks: [count including sub-tasks]

## Task List

### Task Group 1: [Name]

#### Dependencies: [None or previous groups]

- [ ] 1.0 Complete [layer]
  - [ ] 1.1 [Implementation task]
    - [Specific details]
    - [Reuse pattern from: [existing code] if applicable]
  - [ ] 1.2 [Implementation task]
    - [Specific details]
  - [ ] 1.3 [Implementation task]
  - [ ] 1.4 [Implementation task]

**Acceptance Criteria:**
- [Functional criteria]
- [Technical criteria]
- [Quality criteria]

---

### Task Group 2: [Name]

#### Dependencies: Task Group 1

[Similar structure for each group]

---

## Execution Order

Recommended sequence:
1. Task Group 1: [Name] (Foundation)
2. Task Group 2: [Name] (Builds on 1)
3. Task Group 3: [Name] (Builds on 2)
4. Task Group 4: [Name] (Verification)

## Notes

- Tasks reference specific files/components from spec.md
- Reusability opportunities noted where applicable
- Visual assets referenced in frontend tasks
- Follow TDD: write tests for each component/behavior
EOF

Step 5: Report Completion

✅ Task breakdown complete!

Created: tasks.md

Summary:
- [X] task groups by dependencies
- [Y] total tasks
- Execution order defined
- Acceptance criteria for each group

Structure:
- Specific and verifiable tasks
- Dependencies clearly marked
- Reusability notes included
- Visual references added (if applicable)

Ready for Phase 5: Verification

Return to workflow.

Task Breakdown Principles

Grouping Logic

Group by layer AND skill:

  • Database tasks together (backend skills)
  • API tasks together (backend + API design)
  • Frontend tasks together (UI/UX + frontend)
  • Testing together (integration + QA)

Respect dependencies:

  • Foundation layers first
  • Build progressively
  • Verification last

Task Granularity

Good task size: Completable in 1-4 hours

Too granular:

  • ❌ "Add import statement"
  • ❌ "Define variable"

Too broad:

  • ❌ "Build authentication system"
  • ❌ "Create all components"

Just right:

  • ✅ "Create User model with validations"
  • ✅ "Implement login API endpoint"
  • ✅ "Build registration form component"

Testing Approach

Follow TDD naturally:

  • Write test for behavior
  • Implement to pass test
  • Refactor while green
  • Repeat for next behavior

Don't specify test counts - write tests for each:

  • Model method
  • API endpoint
  • UI component behavior
  • Integration flow

Subagents following TDD will write appropriate tests.

Adapting Structure

API-heavy feature:

1. Database Layer
2. API Endpoints
3. Authentication/Authorization
4. Frontend Integration
5. Integration Testing

UI-heavy feature:

1. Data Layer (if needed)
2. Component Library
3. Page Layouts
4. Interaction Flows
5. Integration Testing

Background job:

1. Data Models
2. Job Logic
3. Queue Configuration
4. Monitoring/Logging
5. Integration Testing

Adapt to feature, but always:

  • Respect dependencies
  • Include verification group
  • Note reusability

Red Flags

Never:

  • Create vague tasks without specifics
  • Omit dependencies
  • Skip reusability notes when identified
  • Forget visual references (if visuals exist)

Always:

  • Make tasks specific and verifiable
  • Note dependencies between groups
  • Include reusability references
  • Reference visual files in frontend tasks
  • Define acceptance criteria per group

Integration

Called by:

  • spec-creation-workflow (Phase 4)

Returns to:

  • spec-creation-workflow

Creates:

  • [spec]/tasks.md

Uses:

  • spec.md - Requirements and structure
  • planning/requirements.md - Original requirements
  • planning/visuals/* - Design assets

Next phase uses:

  • tasks.md for verification