- NPX install: npx get-shit-done-cc - Slash commands for project management - Context engineering templates and workflows 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
4.3 KiB
description, allowed-tools
| description | allowed-tools | ||||
|---|---|---|---|---|---|
| Initialize a new project with deep context gathering, PROJECT.md, roadmap, and state tracking |
|
This is the most leveraged moment in any project. Deep questioning here means better plans, better execution, better outcomes.
Creates .planning/ with PROJECT.md, ROADMAP.md, STATE.md, and config.json.
<execution_context>
@/.claude/get-shit-done/workflows/create-roadmap.md
@/.claude/get-shit-done/references/principles.md
@/.claude/get-shit-done/references/questioning.md
@/.claude/get-shit-done/templates/project.md
@~/.claude/get-shit-done/templates/config.json
</execution_context>
# Abort if project exists
[ -d .planning ] && echo "Project already initialized. Use /gsd:progress" && exit 1
# Initialize git
[ -d .git ] || git init
Then use AskUserQuestion to cover the 9 domains (project type, problem, audience, success, constraints, scope, current state, decisions, open questions).
Skip domains already clear from user input. Probe for specifics on vague answers.
Decision gate (MUST have all 3 options):
Header: "Ready?"
Options:
1. "Create PROJECT.md" - Finalize
2. "Ask more questions" - Dig into uncovered domains
3. "Let me add context" - User shares more
If "Ask more questions" → ask about 2-3 uncovered domains → return to gate. Loop until "Create PROJECT.md" selected.
Synthesize all context into `.planning/PROJECT.md` using the template from `templates/project.md`.Do not compress. Capture everything gathered.
Ask workflow mode preference:Use AskUserQuestion:
- header: "Mode"
- question: "How do you want to work?"
- options:
- "Interactive" - Confirm at each step
- "YOLO" - Auto-approve, just execute
Create .planning/config.json with chosen mode using templates/config.json structure.
This handles:
- Domain expertise detection (scan for applicable skills)
- Phase identification (3-6 phases based on PROJECT.md)
- Research needs detection (flag phases needing investigation)
- Phase confirmation (respects yolo/interactive mode)
- ROADMAP.md creation with research flags
- STATE.md initialization
- Phase directory creation
The workflow will create:
.planning/ROADMAP.md.planning/STATE.md.planning/phases/XX-name/directories
[One-liner from PROJECT.md]
Phases:
- EOF )"
</step>
<step name="done">
Project initialized:
- Project: .planning/PROJECT.md
- Roadmap: .planning/ROADMAP.md ([N] phases)
- State: .planning/STATE.md
- Config: .planning/config.json (mode: [chosen mode])
What's next?
- Plan Phase 1 (/gsd:plan-phase 01)
- Review project setup
- Done for now
If user selects "Plan Phase 1" → invoke `/gsd:plan-phase 01`
</step>
</process>
<output>
- `.planning/PROJECT.md`
- `.planning/ROADMAP.md`
- `.planning/STATE.md`
- `.planning/config.json`
- `.planning/phases/XX-name/` directories
</output>
<success_criteria>
- [ ] Deep questioning completed (not rushed)
- [ ] PROJECT.md captures full context
- [ ] config.json has workflow mode
- [ ] ROADMAP.md has 3-6 phases with research flags
- [ ] STATE.md initialized with project summary
- [ ] Phase directories created
- [ ] All committed to git
</success_criteria>