Project Management Commands
Quack provides a powerful project management system through four global slash commands that integrate with the Kanban board and AI droids.
Overview
| Command | Purpose | Use Case |
|---|---|---|
/task | Generic task | Any work item |
/bug | Bug resolution | Fix issues with test coverage |
/feature | New functionality | Create something new |
/improvement | Enhance existing code | Refactoring, performance, UX |
All commands use the quack-pm skill and create tasks on the Kanban board (Cmd+K).
The Core Workflow: Analyze-Plan-Act-Test-Review-Document
All commands follow a structured workflow with 7 phases:
- Analyze - Understand existing code and context
- Plan - Create detailed implementation plan
- Act - Implement the solution
- Test - Write and run tests
- Act - Refine based on test results
- Review - Code review for quality
- Document - Save knowledge for future reference
/task - Generic Tasks
The simplest command for any work item.
Usage:
bash/task Implement dark mode toggle
Workflow:
- Discovery - Find available agents
- Create task on Kanban
- Assign appropriate droid
Best for: Quick tasks, miscellaneous work, one-off items.
/bug - Bug Resolution
Specialized for fixing issues with a focus on root cause analysis and prevention.
Usage:
bash/bug Dropdown menu closes unexpectedly on mobile
Workflow:
- Analyze with
code-explorerdroid - Identify root cause
- Implement fix
- Write test to prevent regression
- Document solution in MCP Memory
Key Features:
- Always starts with code analysis
- Requires test to verify fix
- Automatically saves solution to Second Brain
/feature - New Functionality
For creating something new from scratch.
Usage:
bash/feature Add user preferences panel with theme and language settings
Workflow:
- Analyze -
code-explorerexamines codebase - Plan - Detailed implementation plan
- Act - Implementation with droids:
- Frontend:
frontend-developer - Backend:
data-engineer
- Frontend:
- Test -
test-engineerwrites tests - Review -
code-reviewervalidates - Document - Save to
.claude/docs/
Parallel Execution:
For full-stack features, run droids in parallel:
bash/background @data-engineer Implement API endpoints for user preferences /background @frontend-developer Implement UI for user preferences
Droid Selection:
| Feature Type | Primary Droid |
|---|---|
| UI/React | frontend-developer |
| API/Rust | data-engineer |
| Full-stack | code-explorer (then subtasks) |
/improvement - Enhance Existing Code
For refactoring, performance optimization, UX improvements, and cleanup.
Usage:
bash/improvement Optimize the file tree rendering performance
Key Difference from /feature:
| Aspect | /feature | /improvement |
|---|---|---|
| Code | New | Existing |
| Risk | Low | Medium-High |
| Test first | No | YES |
| Benchmark | Optional | Required (if performance) |
Workflow:
- Analyze deeply - Understand current implementation
- Test FIRST - Verify existing tests pass
- Plan - Incremental refactoring strategy
- Act - Small, safe changes
- Test - Verify no regressions
- Benchmark - Measure improvements
- Review - Validate changes
Improvement Types:
| Type | Focus | Droid |
|---|---|---|
| Refactoring | Code structure | code-explorer |
| Performance | Speed, memory | data-engineer |
| UX | Usability | frontend-developer |
| Cleanup | Tech debt | code-reviewer |
| Security | Vulnerabilities | code-reviewer |
Golden Rule: Existing tests must pass BEFORE and AFTER refactoring!
Droids Reference
The commands automatically assign work to specialized droids:
| Droid | Specialty |
|---|---|
code-explorer | Analyze codebase, find patterns |
frontend-developer | React, UI, state management |
data-engineer | Backend, Rust, APIs, data |
test-engineer | Vitest tests, coverage |
code-reviewer | Code quality, security |
documentation-writer-expert | Technical docs |
second-brain-manager | MCP Memory storage |
Parallel Execution with /background
Speed up work by running droids in parallel:
Full-stack (backend + frontend):
bash/background @data-engineer Implement backend for [feature] /background @frontend-developer Implement UI for [feature]
Test + Review together:
bash/background @test-engineer Write tests for [feature] /background @code-reviewer Review code for [feature]
Documentation (non-blocking):
bash/background @second-brain-manager Save pattern to MCP Memory
When to use parallel:
- Backend/Frontend independent → Parallel
- Tests depend on code → Sequential
- Review + Test → Parallel
- Documentation → Always parallel
Best Practices
Choosing the Right Command
- Something new? →
/feature - Something broken? →
/bug - Something to improve? →
/improvement - Something quick/misc? →
/task
Task Naming
- Use action verbs: "Implement", "Fix", "Optimize", "Add"
- Keep titles short (5-10 words)
- Be specific about scope
Droid Selection
- Start with
code-explorerfor analysis - Match implementation droid to tech stack
- Always use
test-engineerfor tests - End with
code-reviewerfor validation
Documentation
All completed work should be documented:
- Technical docs:
.claude/docs/[name].md - Patterns/solutions: MCP Memory (Second Brain)
- Use
second-brain-managerdroid
Quick Start Example
Creating a new settings feature:
bash# 1. Create the feature task /feature Add settings page with theme toggle and notification preferences # 2. Monitor in Kanban (Cmd+K) # 3. When analysis is done, run implementation in parallel /background @data-engineer Implement settings API endpoints /background @frontend-developer Implement settings UI components # 4. After implementation, run test and review /background @test-engineer Write tests for settings feature /background @code-reviewer Review settings implementation # 5. Document the feature /background @second-brain-manager Document settings feature implementation
See Also
- Kanban Board - Visual task management
- Background Tasks - Non-blocking execution
- Second Brain - Knowledge management