npx claudepluginhub earthmanweb/serena-workflow-engine --plugin serena-workflow-engineWant just this skill?
Then install: npx claudepluginhub u/[userId]/[slug]
Initialize workflow for new empty projects. Creates core memories and directory structure.
This skill uses the workspace's default tool permissions.
⚠️ WORKFLOW INITIALIZATION
If starting a new session, first read workflow initialization:
mcp__plugin_swe_serena__read_memory("wf/WF_INIT")
Follow WF_INIT instructions before executing this skill.
Scaffold Project Skill
Initialize workflow system for new or empty projects.
When to Use
- New projects without existing memories
- Projects missing INDEX_FEATURES
- Converting existing projects to workflow system
Detection Triggers
Automatically suggested when:
- No
.serena/swe/directory exists - No
INDEX_FEATURES.mdfile exists INDEX_FEATURES.mdhas zero features registered
Process
Stage 1: Project Detection
# Detect project root
git rev-parse --show-toplevel || pwd
# Detect package manager
[ -f "package.json" ] && echo "npm"
[ -f "composer.json" ] && echo "composer"
[ -f "Cargo.toml" ] && echo "cargo"
[ -f "requirements.txt" ] && echo "pip"
[ -f "go.mod" ] && echo "go"
# Detect primary language
find . -name "*.ts" -o -name "*.js" | head -1 # TypeScript/JavaScript
find . -name "*.py" | head -1 # Python
find . -name "*.php" | head -1 # PHP
find . -name "*.rs" | head -1 # Rust
find . -name "*.go" | head -1 # Go
Stage 2: Directory Setup
mkdir -p .serena/swe
mkdir -p .claude/skills
mkdir -p .claude/hooks
Stage 3: Core Memory Creation
Create from templates:
- _INDEX - Navigation hub
# _INDEX - Memory Navigation
## Quick Reference
- Features: INDEX_FEATURES
- Architecture: ARCH_INDEX
- Workflows: INDEX_WORKFLOWS_STATES
## Memory Types
| Prefix | Purpose |
| -------- | ----------------- |
| FEATURE_ | Feature configs |
| DOM_ | Domain behaviors |
| SYS_ | System references |
| REF_ | Reference docs |
| INDEX_ | Navigation |
| WF_ | Workflow states |
| WM_ | Session state |
- INDEX_FEATURES - Empty feature registry
# INDEX_FEATURES
## Registered Features
(none yet - run /swe-feature-onboard to add)
## Quick Start
1. `/swe-feature-onboard [KEY]` - Full wizard
2. `/swe-onboard-quick [KEY]` - Fast setup
- ARCH_INDEX - Basic architecture placeholder
# ARCH_INDEX - Architecture Overview
## Project Type
[Detected or unknown]
## Primary Language
[Detected]
## Framework
[Detected or none]
## Structure
(Run /swe-feature-onboard to populate)
Stage 4: First Feature Prompt
PROJECT SCAFFOLDED
Created:
- .serena/swe/
- _INDEX
- INDEX_FEATURES
- ARCH_INDEX
Your project needs at least one feature to enable code changes.
What is the main codebase?
- Name: [e.g., "Backend API"]
- Key: [e.g., "BACKEND"]
- Path: [e.g., "src/"]
Options:
- [A] Set up now with /swe-feature-onboard (recommended)
- [B] Quick setup with /swe-onboard-quick
- [C] Skip - add features later (research-only mode)
Stage 5: Optional Swarm Analysis
If swarm MCP available:
AI-powered codebase analysis available.
[A] Full DAA analysis (creates DOM_*, SYS_*, detailed INDEX_*)
[B] Quick scan (basic structure)
[C] Skip
Minimal Workflow Mode
If user skips feature setup, enable minimal mode:
{
"mode": "minimal",
"allowed_states": ["WF_START", "WF_RESEARCH", "WF_CLARIFY"],
"blocked_states": ["WF_EXECUTE", "WF_CHECKPOINT"],
"message": "Feature onboarding required for code changes"
}
Skill Return Format
## Skill Return
- **Skill**: swe-scaffold-project
- **Status**: [success|needs_clarification]
- **Project Root**: [path]
- **Language**: [detected]
- **Framework**: [detected or none]
- **Memories Created**: _INDEX, INDEX_FEATURES, ARCH_INDEX
- **Next Step Hint**: WF_START or /swe-feature-onboard
Exit
> **Skill /swe-scaffold-project complete** - Project scaffolded, run /swe-feature-onboard to add first feature
Similar Skills
You MUST use this before any creative work - creating features, building components, adding functionality, or modifying behavior. Explores user intent, requirements and design before implementation.