Get Shit Done

A meta-prompting, context engineering and spec-driven development system for Claude Code by TÂCHES.

Vibecoding has a bad reputation. You describe what you want, AI generates code, and you get inconsistent garbage that falls apart at scale.

GSD fixes that. It's the context engineering layer that makes Claude Code reliable. Describe your idea, let the system extract everything it needs to know, and then let Claude Code get to work.

THIS is how you vibecode and actually get shit done.

Warning: Not for people who enjoy inconsistent and sloppy results.

Installation

npx get-shit-done-cc

That's it. Works on Mac, Windows, and Linux.

Verify: /gsd:help

Why I Built This

I'm a solo developer. I don't write code — Claude Code does.

Other spec-driven development tools exist; BMAD, Speckit... But they all seem to make things way more complicated than they need to be (sprint ceremonies, story points, stakeholder syncs, retrospectives, Jira workflows) or lack real big picture understanding of what you're building. I'm not a 50-person software company. I don't want to play enterprise theater. I'm just a creative person trying to build great things that work.

So I built GSD. The complexity is in the system, not in your workflow. Behind the scenes: context engineering, XML prompt formatting, subagent orchestration, state management. What you see: a few commands that just work.

I wanted to spend my time having ideas and seeing them through to implementation — not babysitting Claude. Now I can say "go," put it in YOLO mode, and go to the beach. The system gives Claude everything it needs to do the work and verify it. I trust the workflow. It just does a good job.

That's what this is. No enterprise roleplay bullshit. Just an incredibly effective system for building cool stuff consistently using Claude Code.

— TÂCHES

How It Works

1. Start with an idea

/gsd:new-project

The system asks questions. Keeps asking until it has everything — your goals, constraints, tech preferences, edge cases. You go back and forth until the idea is fully captured. Creates PROJECT.md.

2. Create roadmap

/gsd:create-roadmap     # Create phases and state tracking

Roadmap creation produces:

• ROADMAP.md - Phases from start to finish
• STATE.md - Living memory that persists across sessions

3. Plan and execute phases

/gsd:plan-phase 1      # System creates atomic task plans
/gsd:execute-plan      # Subagent implements autonomously

Each phase breaks into 2-3 atomic tasks. Each task runs in a fresh subagent context — 200k tokens purely for implementation, zero degradation.

4. Ship and iterate

/gsd:complete-milestone   # Archive v1, prep for v2
/gsd:add-phase            # Append new work
/gsd:insert-phase 2       # Slip urgent work between phases

Ship your MVP in a day. Add features. Insert hotfixes. The system stays modular — you're never stuck.

Existing Projects (Brownfield)

Already have code? Start here instead.

1. Map the codebase

/gsd:map-codebase

Spawns parallel agents to analyze your code. Creates .planning/codebase/ with 7 documents:

• STACK.md — Languages, frameworks, dependencies
• ARCHITECTURE.md — Patterns, layers, data flow
• STRUCTURE.md — Directory layout, where things live
• CONVENTIONS.md — Code style, naming patterns
• TESTING.md — Test framework, patterns
• INTEGRATIONS.md — External services, APIs
• CONCERNS.md — Tech debt, known issues, fragile areas

2. Initialize project

/gsd:new-project

Same as greenfield, but the system knows your codebase. Questions focus on what you're adding/changing, not starting from scratch.

3. Continue as normal

From here, it's the same: /gsd:create-roadmap → /gsd:plan-phase → /gsd:execute-plan

The codebase docs load automatically during planning. Claude knows your patterns, conventions, and where to put things.

Why It Works

Context Engineering

Claude Code is incredibly powerful if you give it the context it needs. Most people don't.

GSD handles it for you:

Size limits based on where Claude's quality degrades. Stay under, get consistent excellence.

XML Prompt Formatting

Every plan is structured XML optimized for Claude:

<task type="auto">
  <name>Create login endpoint</name>
  <files>src/app/api/auth/login/route.ts</files>
  <action>
    Use jose for JWT (not jsonwebtoken - CommonJS issues).
    Validate credentials against users table.
    Return httpOnly cookie on success.
  </action>
  <verify>curl -X POST localhost:3000/api/auth/login returns 200 + Set-Cookie</verify>
  <done>Valid credentials return cookie, invalid return 401</done>
</task>

Precise instructions. No guessing. Verification built in.

Subagent Execution

As Claude fills its context window, quality degrades. You've seen it: "Due to context limits, I'll be more concise now." That "concision" is code for cutting corners.

GSD prevents this. Each plan is maximum 3 tasks. Each plan runs in a fresh subagent — 200k tokens purely for implementation, zero accumulated garbage.

• Task 1: fresh context, full quality
• Task 2: fresh context, full quality
• Task 3: fresh context, full quality

No degradation. Walk away, come back to completed work.

Clean Git History

Every task: atomic commit, clear message, summary documenting outcomes. Maintainable history you can trace.

Modular by Design

• Add phases to current milestone
• Insert urgent work between phases
• Complete milestones and start fresh
• Adjust plans without rebuilding everything

You're never locked in. The system adapts.

Commands

Who This Is For

People who want to vibecode and have it actually work.

Anyone who wants to clearly describe what they want, trust the system to build it, and go live their life.

Not for people who enjoy inconsistent and sloppy results.

License

MIT License. See LICENSE for details.

Claude Code is powerful. GSD gives it the context and the systematic consistency to prove it.