GSD (Get Shit Done) is one of the most popular AI coding workflows today, with roughly 64K+ stars on GitHub. It is a spec-driven development framework designed for AI coding tools such as Claude Code, OpenCode, and Gemini CLI.
Core problem it solves: Context Rot
- In conventional AI coding, as a conversation progresses, the context window fills up and AI output quality gradually degrades.
- GSD solves this by giving each task an independent context—every task runs in a fresh 200K-token environment.
- 28+ slash commands
- 12 dedicated agents (planning, execution, verification, research, etc.)
- Pure Markdown files + Node.js CLI, no proprietary runtime
- Wave-based parallelism: analyzes dependencies; independent tasks run in parallel
- Quick mode
/gsd:quick: skips the full workflow, suited to small tasks - PROJECT.md – project vision
- REQUIREMENTS.md – requirements with REQ-IDs
- ROADMAP.md – phases and success criteria
- PLAN.md – atomic task plans (XML structure)
- STATE.md – project state tracking
- GitHub: https://github.com/gsd-build/get-shit-done
- Install:
npx get-shit-done-cc - Docs: https://zread.ai/gsd-build/get-shit-done
Architecture:
---
Core Workflow (Five Stages)
| Stage | Command | Purpose |
|-----|------|-----|
| Init | /gsd:new-project | Deep questioning → domain research → requirements extraction → roadmap generation |
| Discuss | /gsd:discuss-phase N | Clarify implementation preferences, visual style, API design, etc. |
| Plan | /gsd:plan-phase N | Research implementation approach, generate a plan of 2-3 atomic tasks |
| Execute | /gsd:execute-phase N | Execute plans in parallel; each task in an independent context with atomic commits |
| Verify | /gsd:verify-work N | Confirm goals are achieved, not just tasks completed |
Execution modes:
---
Common Commands
| Command | Use |
|------|------|
| /gsd:new-project [--auto] | Create a new project; auto mode skips interaction |
| /gsd:map-codebase | Analyze an existing codebase (legacy projects) |
| /gsd:discuss-phase N | Discuss implementation details of phase N |
| /gsd:plan-phase N [--prd] | Create an execution plan for phase N |
| /gsd:execute-phase N [--gaps-only] | Execute the phase N plan |
| /gsd:verify-work N | Verify phase N work |
| /gsd:quick [--discuss] [--full] | Quick mode for small tasks |
| /gsd:complete-milestone | Archive, tag, and initialize the next cycle |
---
Key Documents Generated
---
Pros and Cons
✅ Pros
1. Solves context rot – each task gets a fresh 200K context; stable quality 2. Atomic commits – each task gets its own commit; clean git history 3. Parallel execution – independent tasks can run in parallel, boosting efficiency 4. Quality verification – every phase includes verification to ensure goals are met 5. No vendor lock-in – plain Markdown + standard Claude Code features 6. Industry recognition – used by engineers at Amazon, Google, Shopify, Webflow, etc.
❌ Cons
1. Not fully automatic – each stage boundary must be manually triggered by a command 2. Learning curve – requires adopting the spec-driven development mindset 3. Large upfront investment – initialization involves answering many questions 4. Not for exploratory coding – better suited to projects with clear goals 5. Claude Code dependency – the ecosystem mainly lives in Claude Code
---
TL;DR
GSD is a framework that turns chat into engineering—trading some of the fluidity of real-time interaction for a predictable, verifiable, and scalable AI coding experience. It fits scenarios that deliver production-grade code rather than rapid prototyping.
---