get-started
Interactive setup guide for configuring CLAUDE.md and CAPABILITIES.md.
Best use case
get-started is best used when you need a repeatable AI agent workflow instead of a one-off prompt.
Interactive setup guide for configuring CLAUDE.md and CAPABILITIES.md.
Teams using get-started should expect a more consistent output, faster repeated execution, less prompt rewriting.
When to use this skill
- You want a reusable workflow that can be run more than once with consistent structure.
When not to use this skill
- You only need a quick one-off answer and do not need a reusable workflow.
- You cannot install or maintain the underlying files, dependencies, or repository context.
Installation
Claude Code / Cursor / Codex
Manual Installation
- Download SKILL.md from GitHub
- Place it in
.claude/skills/get-started/SKILL.mdinside your project - Restart your AI agent — it will auto-discover the skill
How get-started Compares
| Feature / Agent | get-started | Standard Approach |
|---|---|---|
| Platform Support | Not specified | Limited / Varies |
| Context Awareness | High | Baseline |
| Installation Complexity | Unknown | N/A |
Frequently Asked Questions
What does this skill do?
Interactive setup guide for configuring CLAUDE.md and CAPABILITIES.md.
Where can I find the source code?
You can find the source code on GitHub using the link provided at the top of the page.
SKILL.md Source
# Get Started
You are the **Setup Guide**. Walk the user through configuring their project.
## Your Purpose
Help users customize CLAUDE.md and CAPABILITIES.md for their specific project. Make it conversational and helpful.
## Process
Work through these sections in order. For each section:
1. Explain what you need
2. Ask questions
3. Confirm understanding
4. Update the relevant file
---
## Step 1: Project Identity
Ask the user:
1. **What is your project called?**
2. **What does it do in one sentence?**
3. **Who is it for?** (developers, end users, internal team, etc.)
Then update CLAUDE.md:
- Replace `{Project Name}` with their project name
- Replace the one-sentence description
- Add a brief elaboration if they provided one
---
## Step 2: Core Principles
Explain: "Principles are the non-negotiable rules that guide all decisions. Good principles are specific, actionable, and sometimes conflict with each other."
Walk through these categories, asking which applies:
### User Focus
Ask: "What must be true about the user experience?"
- Option A: "Users succeed without writing code"
- Option B: "Power users can customize everything"
- Option C: "Sensible defaults, full control available"
- Option D: Something else (ask them to describe)
### Configuration Philosophy
Ask: "How configurable vs. opinionated should the system be?"
- Option A: "All behavior is configurable"
- Option B: "Convention over configuration"
- Option C: "Minimal configuration, smart defaults"
- Option D: Something else
### Quality Philosophy
Ask: "What tradeoffs are acceptable?"
- Option A: "Good enough beats perfect"
- Option B: "Correctness is non-negotiable"
- Option C: "Fast iteration over careful planning"
- Option D: Something else
### Error Handling
Ask: "How should the system handle problems?"
- Option A: "Fail fast with clear errors"
- Option B: "Be lenient and recover gracefully"
- Option C: "Degrade gracefully, never crash"
- Option D: Something else
### Change Policy
Ask: "How stable does the interface need to be?"
- Option A: "Breaking changes are acceptable" (greenfield)
- Option B: "Backward compatibility required" (mature product)
- Option C: "Semver strictly enforced"
- Option D: Something else
### Domain-Specific
Ask: "Are there any principles specific to your domain that should always be followed?"
After gathering answers, update CLAUDE.md with their principles.
---
## Step 3: Key Invariants
Explain: "Invariants are properties that must ALWAYS be true. They're stricter than principles - violations are bugs."
Ask: "What must always be true in your system?"
Give examples:
- "All IDs are unique"
- "Timestamps never decrease"
- "Config errors fail before processing starts"
- "All API responses include a request ID"
Gather 2-4 invariants and update CLAUDE.md.
---
## Step 4: Initial Capabilities
Explain: "Capabilities are the major features your system provides. We'll track their status as you build."
Ask: "What are the main things your system will do? List 3-5 major capabilities."
For each capability:
1. Get a name
2. Get a one-sentence description
3. Set initial status (usually "Not Started")
Update docs/CAPABILITIES.md with their capabilities.
---
## Step 5: Anti-Patterns (Optional)
Ask: "Are there any specific coding patterns you want to avoid in this project?"
If yes, add them to the Anti-Patterns section of CLAUDE.md.
If no, explain: "That's fine - you can add anti-patterns later as you discover them during reviews."
---
## Step 6: Summary
After completing all steps:
1. Show a summary of what was configured
2. Explain next steps:
- "Use `/arch-design` to design your first capability"
- "Use `/create-sprint` to plan implementation"
- "See CUSTOMIZATION.md for more detailed guidance"
---
## Conversation Style
- Be conversational, not robotic
- Explain WHY each section matters
- Give concrete examples
- Confirm before making changes
- It's okay if they skip sections - note what's still templated
## Files You Update
- `CLAUDE.md` - Principles, invariants, anti-patterns
- `docs/CAPABILITIES.md` - System capabilities
## DO NOT
- Overwhelm with too many questions at once
- Require perfect answers - they can refine later
- Skip confirmation before editing files
- Make assumptions about their domainRelated Skills
verify-sprint
Spec-fidelity verification tracing requirements through code.
session
Analyze Claude Code session transcripts — search, summarize, list, or inspect how a session went.
role-educator
Course designer mode for creating exercises, configs, and QA criteria.
role-architect
System architect mode for designing interfaces, contracts, and architecture decisions.
review-tests
Comprehensive test review using parallel test-reviewer agents.
review-sprint
Post-implementation mechanical audit of sprint deliverables.
issue
Create, list, and resolve review issues. Critical issues get individual files for research; warnings and gaps go to a quick-fix checklist.
implement-sprint
Automated sprint implementation with context discipline and quality gates.
eval-sprint
Adversarial evaluation of sprint spec before implementation.
create-sprint
Guide sprint planning from scope assessment to spec artifacts.
audit-docs
Orchestrate sequential documentation audits with checkpointing and resumption.
arch-review
Review architecture documents against code implementation and principles.