swe-sync

## ⚠️ WORKFLOW INITIALIZATION

**If starting a new session**, first read workflow initialization:
```
mcp__plugin_swe_serena__read_memory("WF_INIT")
```
Follow WF_INIT instructions before executing this skill.

---

# /swe-sync

Synchronize Serena memories between plugin and local project using ruv-swarm coordination.

## Usage

```
/swe-sync                           # Sync all memories plugin → local
/swe-sync --dry-run                 # Preview changes without syncing
/swe-sync category=wf               # Sync only WF_ workflow files
/swe-sync category=ref              # Sync only REF_ reference files
/swe-sync direction=local-to-plugin # Sync local changes back to plugin
```

## Process

### Step 1: Initialize Swarm

```javascript
// Initialize ruv-swarm with mesh topology
mcp__ruv-swarm__swarm_init({ topology: "mesh", strategy: "balanced", maxAgents: 5 })

// NOTE: daa_init is NOT needed here - we use agent_spawn + task_orchestrate pattern
// DAA is only needed when using daa_agent_create + daa_workflow_execute pattern
```

### Step 2: Spawn Comparison Agents

```javascript
// Spawn agents for parallel comparison
mcp__ruv-swarm__agent_spawn({ type: "analyst", name: "plugin-scanner", capabilities: ["file_analysis"] })
mcp__ruv-swarm__agent_spawn({ type: "analyst", name: "local-scanner", capabilities: ["file_analysis"] })
mcp__ruv-swarm__agent_spawn({ type: "coordinator", name: "diff-reporter", capabilities: ["synthesis"] })
```

### Step 3: Orchestrate Comparison Task

```javascript
mcp__ruv-swarm__task_orchestrate({
  task: "Compare memory files between plugin and local project",
  strategy: "parallel",
  priority: "high"
})
```

### Step 4: Execute File Comparison

**Plugin Path:** `.claude/plugins/serena-workflow-engine/memories/`
**Local Path:** `.serena/memories/`

**Categories:**
| Category | Pattern | Description |
|----------|---------|-------------|
| wf | `wf/WF_*.md` | Workflow state files |
| ref | `ref/REF_*.md` | Reference documentation |
| all | `