Initialize Project
"Start as you mean to go on."
Philosophy
All other skills depend on shared configuration and context. This skill creates that foundation. Run once per project. Takes 2 minutes.
Workflow
Step 1: Ask Configuration Questions
Use defaults if user doesn't want to decide:
- AI Tools — Which tools? (Claude Code, Cursor, Windsurf, Cline, Copilot, Codex, Aider, Gemini CLI, OpenCode, Copilot CLI, Codex CLI, Other)
- Issue Tracker — GitHub Issues (default) / Linear / Local Files
- Triage Labels —
bug,feature,enhancement,tech-debt,question - Docs Path —
docs/(default),docs/adr/,docs/handoff/ - Context File —
CONTEXT.mdat project root (default) - Test Framework — What framework? (jest, pytest, go test, etc.)
- Code Style — Project-specific rules?
Step 2: Create Files
Always create:
CONTEXT.md— Domain context and shared languagedocs/adr/0001-initial-setup.md— Initial ADR
Tool-specific (based on Step 1):
| Tool | File |
|---|---|
| Claude Code | CLAUDE.md |
| Cursor | .cursor/rules/integrated-guidelines.mdc |
| Windsurf | .windsurfrules |
| Cline | .clinerules |
| GitHub Copilot | .github/copilot-instructions.md |
| Codex | AGENTS.md |
| Aider | .aider.conf.yml |
| Gemini CLI | GEMINI.md |
| OpenCode | .opencode/instructions.md |
| Copilot CLI | .copilot-cli/instructions.md |
| Codex CLI | .codex/instructions.md |
| Other | SYSTEM_PROMPT.md (copy content to tool's custom instructions) |
Step 3: Write CONTEXT.md
# Project Context
## Domain Language
- [Term 1]: [definition]
- [Term 2]: [definition]
## Module Map
- [Module A]: [responsibility]
- [Module B]: [responsibility]
## Architecture Decisions
- See `docs/adr/` for all decisions
## Code Style
- [Rule 1]
- [Rule 2]
Step 4: Write Initial ADR
# ADR 0001: Initial Project Setup
## Status
Accepted
## Context
Setting up AI Engineering Mastery skills for this project.
## Decision
- Issue tracker: [choice]
- AI tools: [choices]
- Test framework: [choice]
## Consequences
- All skills will use this configuration
- CONTEXT.md is the source of truth for domain language
Step 5: Output Summary
Setup Complete!
Configuration:
- Issue Tracker: [choice]
- AI Tools: [choices]
- Docs Path: [path]
- Test Framework: [choice]
Files Created:
- CONTEXT.md
- docs/adr/0001-initial-setup.md
- [tool-specific files]
Available Skills:
/brainstorm — Design exploration with HARD-GATE
/grill — Deep alignment before coding
/tdd — Test-driven development
/diagnose — Systematic debugging
/architect — Architecture improvement
/zoom-out — System-level context
/to-prd — Create PRD
/to-issues — Break plan into issues
/caveman — Compressed communication
/handoff — Session handoff
/verify — Evidence-based completion verification
/review — Code review (request + receive)
Quick Start:
1. Start with /grill to align on your first task
2. Use /tdd to implement with confidence
3. End with /handoff to save your progress
Anti-Patterns
- Over-configuring — Use defaults, adjust later
- Skipping setup — Other skills depend on this
- Not updating CONTEXT.md — It should evolve with the project
Integration
- All other skills depend on this configuration
- Run
/setupbefore any other skill - Re-run if configuration changes significantly
Source: novvoo/ai-engineering-mastery — distributed by TomeVault.