Get Shit Done (GSD)
Overview
This skill provides a comprehensive workflow for executing large and complex projects with minimal chaos.
It focuses on breaking projects into manageable steps, maintaining documentation, and reducing debugging through structured execution.
Setup
Before using this skill:
- Define the project goal
- Create a workspace for:
- documentation
- task tracking
- progress logs
Recommended tools:
- GitHub Projects
- Markdown documentation
- Kanban boards
- Issue tracking systems
Inputs Required
- Project idea or objective
- Requirements or constraints
- Timeline or priorities (optional)
When to Use This Skill
- Large software projects
- Multi-step workflows
- Long-term execution planning
- Complex implementation tasks
- Team coordination
When NOT to Use
- Tiny one-step tasks
- Quick experiments
- Highly exploratory brainstorming sessions
Example Use Case
“Build a full-stack AI application with authentication and deployment”
Claude should:
- Break project into phases
- Create implementation roadmap
- Define milestones
- Generate task-by-task execution plan
- Track progress continuously
Workflow
1. Project Definition
- Define end goal
- Clarify constraints
- Identify major components
2. Task Decomposition
Break project into:
- modules
- milestones
- subtasks
- dependencies
3. Execution Planning
- Prioritize tasks
- Estimate complexity
- Define deliverables
4. Implementation
- Execute incrementally
- Validate after each stage
- Document decisions
5. Tracking & Iteration
- Monitor progress
- Detect blockers early
- Refine plans dynamically
Output Expectations
- Structured execution roadmap
- Clear task hierarchy
- Reduced implementation chaos
- Better debugging and maintainability
Execution Strategy (for AI agents)
The agent should:
- Prevent scope chaos
- Break large problems into manageable tasks
- Maintain structured documentation
- Track dependencies continuously
- Validate progress incrementally
Best Practices
- Avoid implementing everything at once
- Document decisions early
- Validate small increments frequently
- Keep tasks independently testable
Notes
- Large projects fail primarily from poor structure
- Incremental execution reduces debugging overhead
- Good documentation improves long-term maintainability