Commands Reference

This is the narrated, namespace-grouped reference (126 core commands). For the complete auto-generated table of all 220 commands (core + infra, with source links), see COMMANDS-FULL-REFERENCE.md. For the claude-craft NPX CLI (a different tool — install, init, check…), see CLI-REFERENCE.md.

Claude Code commands are slash commands that automate workflows and provide structured assistance.

How to Use Commands

Type a slash command in Claude Code:

/common:pre-commit-check
/symfony:generate-crud User
/react:generate-component Button

Commands can take arguments:

/command:name argument1 argument2

Plan Mode Classification

Every command includes a Plan Mode section that indicates when Claude should activate plan mode before executing. This ensures proper analysis and validation for impactful operations.

Level	Description	Example Commands
MANDATORY	Plan mode activates automatically before execution. Claude analyzes impacted code and proposes a plan for your validation.	/workflow:implement, /qa:tdd, /qa:recette, generate-* commands
RECOMMENDED	Plan mode is recommended for complex scenarios. Activates when scope spans multiple modules.	/workflow:plan, /workflow:design, /common:architecture-decision
CONDITIONAL	Plan mode activates automatically when scope is broad (multiple modules, cross-cutting analysis).	check-* commands, /workflow:analyze, /common:research-context7

Tip: Commands without a Plan Mode section (e.g., /workflow:status, /common:daily-standup) are lightweight and never require plan mode.

---

Command Namespaces

Les 15 namespaces cœur + technologies (/common, /workflow, /team, /qa, /uiux + les 11 stacks dont /paperclip:*) totalisent 126 commandes — c'est le chiffre d'en-tête de Claude Craft. Le tableau ci-dessous inclut en plus les namespaces d'infrastructure et de gestion de projet (BMAD), pour un total installable de 220 commandes sur 27 namespaces. Référence détaillée auto-générée : COMMANDS-FULL-REFERENCE.

Namespace	Technology	Count
/common:	Transversal	19
/workflow:	Workflow (BMAD)	10
/team:	Agent Teams	4
/qa:	QA & Testing	6
/uiux:	UI/UX & Accessibility	8
/symfony:	PHP/Symfony	10
/flutter:	Dart/Flutter	10
/python:	Python	10
/react:	React/TypeScript	10
/reactnative:	React Native	10
/angular:	Angular	6
/csharp:	C#/.NET	6
/laravel:	PHP/Laravel	6
/vuejs:	Vue.js	6
/php:	PHP	5
/docker:	Docker/Infrastructure	5
/coolify:	Coolify/PaaS	5
/kubernetes:	Kubernetes/Infrastructure	5
/opentofu:	OpenTofu/IaC	5
/ansible:	Ansible/Automation	5
/hcloud:	Hcloud/Hetzner Cloud	5
/pgbouncer:	PgBouncer/Connection Pooling	5
/frankenphp:	FrankenPHP/PHP Server	5
/paperclip:	Paperclip AI-workforce orchestration	8
/sprint:	Sprint Management (BMAD v6)	5
/gate:	Quality Gates (BMAD v6)	7

---

Command Reference by Namespace

Each namespace has its own focused page:

Namespace	Reference
Common Commands (/common:)	View commands
Workflow Commands (/workflow:)	View commands
Symfony Commands (/symfony:)	View commands
Flutter Commands (/flutter:)	View commands
Python Commands (/python:)	View commands
React Commands (/react:)	View commands
React Native Commands (/reactnative:)	View commands
Angular Commands (/angular:)	View commands
C#/.NET Commands (/csharp:)	View commands
Laravel Commands (/laravel:)	View commands
Vue.js Commands (/vuejs:)	View commands
PHP Commands (/php:)	View commands
Docker Commands (/docker:)	View commands
Coolify Commands (/coolify:)	View commands
Kubernetes Commands (/kubernetes:)	View commands
OpenTofu Commands (/opentofu:)	View commands
Ansible Commands (/ansible:)	View commands
Hcloud Commands (/hcloud:)	View commands
PgBouncer Commands (/pgbouncer:)	View commands
FrankenPHP Commands (/frankenphp:)	View commands
Paperclip Commands (/paperclip:)	View commands
Project Commands (/project:)	View commands
Sprint Management Commands (/sprint:)	View commands
Quality Gate Commands (/gate:)	View commands
QA Recette Commands (/common:)	View commands

Command Output Formats

Commands typically output in structured formats:

Audit Commands

══════════════════════════════════════════════════════════════
AUDIT REPORT
══════════════════════════════════════════════════════════════

[ ] Issue 1
[x] Check passed
[ ] Issue 2

Summary: 2 issues found

Generation Commands

Generated files:
- src/Component/NewComponent.tsx
- src/Component/NewComponent.test.tsx
- src/Component/NewComponent.stories.tsx

Check Commands

Architecture Check
──────────────────────────────────────────────────────────────
✓ Layer separation: PASS
✗ Dependency direction: FAIL
  - Infrastructure depends on Domain (src/Infra/Service.php:15)

Score: 85/100

---

Command Frontmatter Format

All commands include YAML frontmatter for Claude Code discovery:

---
description: Brief description of what the command does
argument-hint: <required-arg> [optional-arg]
---

# Command Title

Command content...

Frontmatter Fields

Field	Required	Description
description	Yes	Brief description shown in command list
argument-hint	No	Shows expected arguments format

Examples

Command with arguments:

---
description: Generate CRUD operations for an entity with Clean Architecture
argument-hint: <EntityName>
---

# Generate CRUD

Generate complete CRUD operations for $ARGUMENTS...

Command without arguments:

---
description: Run all pre-commit checks (tests, lint, security)
---

# Pre-Commit Check

Execute all validation checks before committing...

Variable Substitution

Commands can use these variables:

Variable	Description
$ARGUMENTS	All arguments passed after command name
$ARG1, $ARG2...	Individual positional arguments

---

Creating Custom Commands

Add markdown files to .claude/commands/{namespace}/:

---
description: What this command does
argument-hint: <arg1> [arg2]
---

# My Custom Command

Description of what this command does.

## Arguments
$ARGUMENTS

## Process

### Step 1
What to do first...

### Step 2
What to do next...

## Output
Expected output format...

Place in:

• .claude/commands/common/ for /common: namespace
• .claude/commands/myproject/ for /myproject: namespace

---

Context Management Best Practices

Effective context management is the #1 productivity factor with Claude Code (source: Anthropic).

Key Principles

Principle	Action
Context is finite	Monitor context % in status line
Use /clear	Between unrelated tasks
Delegate investigations	Use sub-agents (Task tool) for exploration
Verification loops	Always provide tests/expected outputs
Plan before acting	Use Plan Mode for complex tasks (> 3 files)
Parallel worktrees	git worktree add for concurrent sessions

CLAUDE.md Size

Keep your main CLAUDE.md under 200 lines. Each additional instruction dilutes attention on existing ones. Use .claude/rules/ for detailed guidelines.

Hooks vs Instructions

Mechanism	Strength	Use When
CLAUDE.md	Suggestion	Guidelines, conventions
Rules (.claude/rules/)	Strong suggestion	Detailed best practices
Hooks	Enforcement	Security constraints, formatting

See .claude/templates/hooks/ for ready-to-use hook templates.

Context Thresholds

Context %	Recommended Action
< 30%	Normal operation
30-60%	Monitor, avoid unnecessary file reads
60-80%	Delegate to sub-agents, consider /clear
> 80%	Compaction imminent, save critical context

See rule 12-context-management.md for full documentation.

---

Best Practices

1. Use check commands before commits
2. Generate with commands for consistent code
3. Audit regularly with team-audit
4. Document decisions with architecture-decision
5. Track sprints with sprint commands
6. Manage context — use /clear between tasks, delegate investigations
7. Use hooks for security enforcement (not just CLAUDE.md instructions)