@realblxckcodex/opencode-swarm 0.6.0

package/CHANGELOG.md

@@ -0,0 +1,560 @@
+# Changelog
+
+> **Languages:** 🇩🇪 [Deutsch](#deutsch) | 🇬🇧 [English](#english)
+
+---
+
+<details>
+<summary>🇩🇪 <strong>Deutsch</strong> | German</summary>
+
+<a name="deutsch"></a>
+
+## 0.6.0 - 2026-02-28 - 🎉 TRUE SUBAGENT BREAKTHROUGH
+
+### Phase 6: True Subagent Orchestration ✅
+
+**Das fehlende Puzzle-Teil ist gefunden!** Nach Analyse von subtask2 haben wir die echte Subtask-Implementierung übernommen.
+
+#### Phase 6.1: True Subtask Parts (`type: "subtask"`)
+- **SwarmManager vollständig überarbeitet** (`swarm-manager.ts`)
+  - Nutzt jetzt `client.session.promptAsync()` mit `type: "subtask"` Parts
+  - Jeder Worker ist als eigener Subtask im Chat sichtbar
+  - Echte Session-Isolation für jeden Worker
+  - Named Result Capture mit `as: "name"` und `$RESULT[name]`
+
+**Vorher vs Nachher:**
+```typescript
+// VORHER (nicht sichtbar):
+await session.prompt({ sessionID, prompt, agent: "coder" })
+
+// NACHHER (eigener Subtask im Chat):
+await client.session.promptAsync({
+  path: { id: sessionID },
+  body: {
+    parts: [{
+      type: "subtask",
+      agent: "coder",
+      model: modelSelection,
+      description: "Worker: Implement feature",
+      prompt: "...",
+      as: "worker_1"
+    }]
+  }
+})
+```
+
+#### Phase 6.2: Swarm Master Agent
+- **Neuer Default Agent** (`agents/swarm-master.md`)
+  - Automatische Erkennung wann Swarm nötig ist
+  - Einfache Tasks → Direkte Antwort
+  - Komplexe Tasks → Automatischer `/swarm_run` Aufruf
+  - Intelligente Task-Klassifizierung
+
+#### Phase 6.3: Complete Agent Army
+- **14 spezialisierte Agenten** erstellt:
+  - **Queens:** strategic-queen, tactical-queen, adaptive-queen
+  - **Workers:** researcher, coder, tester, analyst, architect, reviewer, optimizer, documenter
+  - **Swarm-Spezialisten:** swarm-planner, swarm-implementer, swarm-reviewer
+  - **Master:** swarm-master (User Interface)
+- Alle konsistent mit .env Modellen konfiguriert
+- Jeder Agent mit spezifischem Model, Tools und Instructions
+
+#### Phase 6.4: Subtask Parsing & Role Detection
+- **`parseSubtasksFromPlan()`** - Extrahiert nummerierte Listen aus Planner-Output
+- **`detectRoleFromTask()`** - Weist automatisch Rollen basierend auf Task-Inhalt zu
+- Dynamische Worker-Zuweisung (nicht mehr statisch)
+
+#### Phase 6.5: Comprehensive Logging System
+- **Logger** (`logger.ts`) mit 5 Levels: error, warn, info, debug, trace
+- Component-basierte Logger für detaillierte Debugging
+- **`swarm_logs` Tool** - Abrufen und Filtern von Logs
+- Vollständige Transparenz bei Subtask-Ausführung
+
+### Erfolgsnachweis
+
+**Test-Ergebnis:**
+```
+User: /swarm_run task="Analyze the swarm plugin architecture"
+
+Output:
+✅ Strategic-Queen erstellt Plan (5 Subtasks) - SICHTBAR als Subtask
+✅ Swarm-Implementer führt Analyse durch - SICHTBAR als Subtask  
+✅ 764-Zeilen Architektur-Report generiert
+✅ Alle Komponenten dokumentiert
+```
+
+### Technical Details
+
+**Key Insight aus subtask2:**
+- OpenCode's `task` Tool zeigt keine Subtasks an
+- `type: "subtask"` Parts erzeugen echte Subtask-Sessions
+- Dies ist undokumentiert aber funktioniert perfekt
+
+**Implementierte Patterns:**
+- Result Capture: `{as: "name"}` → `$RESULT[name]`
+- Parallel Execution: Alle Worker gleichzeitig
+- Return Chains: Post-execution Actions
+- Progress Reporting: Live-Updates während Ausführung
+
+#### Neue Files
+- `.opencode/swarm/core/logger.ts` (170 Zeilen)
+- `.opencode/swarm/core/swarm-manager.ts` (vollständig überarbeitet, 450 Zeilen)
+- `.opencode/plugins/swarm/tools/swarm-logs.ts` (54 Zeilen)
+- `.opencode/agents/swarm-master.md` (90 Zeilen)
+- Alle 14 Agent-Definitionen
+
+---
+
+## 0.5.0 - 2026-02-28
+
+### Phase 5: Advanced Features ✅
+
+#### Phase 5.1: MCP-Server Integration
+- **MCP Types** (`mcp.ts`) - MCP Protokoll Typen
+- **MCP Server** (`createMCPServer()`) - Ressourcen-, Tool- & Prompt-Registrierung
+- **MCP Client** (`createMCPClient()`) - Ressourcen-Zugriff, Tool-Ausführung, Prompt-Abruf
+- **Swarm Integration** (`createSwarmMCPAdapter()`) - Kompatibilität mit Claude Desktop
+
+#### Phase 5.2: Dynamic Agent Discovery
+- **Service Registry** (`discovery.ts`) - Agent-Registrierung und Discovery
+- **Load Balancing** - `selectLeastLoaded()`, `selectRoundRobin()`
+- **Health Monitoring** - Automatische Health-Checks, TTL-basierte Cleanup
+- **Agent Watcher** (`createAgentWatcher()`) - Echtzeit-Updates
+- **Hot Reload** (`createHotReloadManager()`) - Automatisches Reload
+
+#### Phase 5.3: Swarm-Mail System
+- **Mailbox** (`mail.ts`) - Per-Agent Mailbox mit Prioritäten
+- **Mail Features** - Threading, Attachments, Metadata
+- **Mail Server** (`createMailServer()`) - Multi-Agent Routing
+- **Templates** (`createTemplate()`, `renderTemplate()`)
+
+#### Phase 5.4: Dashboard/API
+- **Dashboard API** (`dashboard.ts`) - Aggregierte Statistiken
+- **Control Operations** - Pause/Resume/Cancel/Retry
+- **Export** - Prometheus/JSON Export
+- **REST Server** (`createRESTServer()`)
+- **WebSocket Server** (`createWebSocketServer()`)
+- **HTML Dashboard** (`generateDashboardHTML()`)
+
+---
+
+## 0.4.6 - 2026-02-28
+
+### Phase 4: Production-Readiness ✅
+
+#### Phase 4.1: Erweiterte Observability
+- **Metrics Collector** (`metrics.ts`) - Prometheus-kompatibler Export
+- **Tracer** (`tracer()`) - Distributed Tracing mit Parent-Child Spans
+
+#### Phase 4.2: SQLite Task Queue (In-Memory)
+- **Task Queue** (`queue.ts`) - Prioritäts-basiertes Scheduling
+- **Queue Processor** (`createQueueProcessor()`) - Concurrency-Limiting
+
+#### Phase 4.3: State Recovery & Checkpointing
+- **State Manager** (`recovery.ts`) - JSON-Persistenz
+- **Recovery Coordinator** - Crash-Recovery
+
+#### Phase 4.4: Integration Tests
+- **E2E Test Suite** (`integration.spec.ts`) - 689 Zeilen
+
+#### Phase 4.5: Error Enrichment
+- **Hierarchische Fehlerklassen** (`errors.ts`)
+- **Error Codes** - 25+ spezifische Codes
+- **Recovery Suggestions** - Automatische Vorschläge
+- **Error Factory** - `ErrorFactory.sessionNotFound()`, etc.
+- **Retry Decorator** - `withRetry()` mit Exponential Backoff
+
+---
+
+## 0.4.5 - 2026-02-28
+
+### Phase 3: Echte Subagenten-Implementierung ✅
+
+#### Phase 3.1: Task/Subagent Dispatcher
+- **Dispatcher** (`dispatcher.ts`) mit 3 Strategien
+- Native → Fork → Prompt graceful degradation
+
+#### Phase 3.2: Worker-Isolation
+- **Worker Factory** (`worker.ts`) - Isolierte Worker-Erstellung
+- **Worker Pool** - Queue und Concurrency-Limiting
+
+#### Phase 3.3: Message-Bus
+- **Message Bus** (`messaging.ts`) - P2P, Broadcast, Request/Response
+
+#### Phase 3.4: Context Propagation
+- **Context Manager** (`context.ts`) - Parent-Child Hierarchie
+- **Checkpoint Manager** - Automatische Checkpoints
+
+---
+
+## 0.4.4 - 2026-02-28
+
+### Phase 2.5: Workflow Engine (subtask2-Konzepte) ✅
+
+- **Complete Workflow Engine** - Sequential, Parallel, Hierarchical
+- **Result Store** (`results.ts`) - `$RESULT[name]` Variable Resolution
+- **Workflow Executor** (`executor.ts`) - Step Execution mit Retry
+- **Loop Engine** (`loop.ts`) - LLM-basierte Loop Evaluation
+- **Parallel Coordinator** (`parallel.ts`) - Dependency Graph, Topological Sort
+
+---
+
+## 0.4.3 - 2026-02-28
+
+### Phase 2.4: Adapter-Interface vereinheitlichen ✅
+
+- **Vereinheitlichtes Adapter-Interface** (`base.ts`)
+- `BaseSessionAdapter` Interface mit Required/Optional Methoden
+- `AdapterError` Klasse mit Error-Codes
+- Retry-Logic mit Exponential Backoff
+
+---
+
+## 0.4.2 - 2026-02-28
+
+### Phase 2.3: Agent-Registry als eigenes Modul ✅
+
+- **Agent Registry** (`registry.ts`) - CRUD mit Capability-Erkennung
+- Neue Tools: `swarm_agent_list`, `swarm_agent_find`, `swarm_agent_get`, `swarm_agent_stats`
+
+---
+
+## 0.4.1 - 2026-02-28
+
+### Phase 2.2: Config-System mit Zod-Validation ✅
+
+- **Erweitertes Config-System** - File-Support und Hot-Reload
+- `createConfigManager()` mit `.opencode/swarm.config.json`
+- Env-Validation mit klaren Fehlermeldungen
+
+---
+
+## 0.4.0 - 2026-02-28
+
+### Phase 2.1: Plugin Modularisierung ✅
+
+- **Komplette Architektur-Überarbeitung** (950 → ~580 Zeilen, 12 Module)
+- Neue Struktur: `swarm/{types,constants,config,utils,tools,hooks}/`
+
+---
+
+## 0.3.1 - 2026-02-28
+
+### Hotfix
+
+- **ProviderModelNotFoundError behoben**
+- Models werden nur gesetzt wenn definiert
+
+---
+
+## 0.3.0 - 2026-02-28
+
+### Phase 1: Kritische Stabilitäts-Fixes ✅
+
+- **Binding-Bug behoben** (`this._client is undefined`)
+- **Concurrency-Limiter** (max 3 parallele Worker)
+- **Retry-Logic** mit Exponential Backoff
+
+---
+
+## 0.2.0 - 2026-02-28
+
+- Coordinator guard enforcement mit dokumentierten Hooks
+- Compaction-resume context injection
+- Diagnostic tools: `swarm_trace`, `swarm_guard_status`
+- Child-session isolation für `swarm_run`
+- Prompt timeout handling
+- Role-based model mapping via env
+
+---
+
+## 0.1.0 - 2026-02-27
+
+- Initial production-focused Swarm plugin baseline
+- SDK-first session orchestration
+- Security hooks für tool execution
+- Persistent JSON memory
+- Session utility tools
+
+</details>
+
+---
+
+<details open>
+<summary>🇬🇧 <strong>English</strong> | Englisch</summary>
+
+<a name="english"></a>
+
+## 0.6.0 - 2026-02-28 - 🎉 TRUE SUBAGENT BREAKTHROUGH
+
+### Phase 6: True Subagent Orchestration ✅
+
+**The missing puzzle piece is found!** After analyzing subtask2, we adopted the true subtask implementation.
+
+#### Phase 6.1: True Subtask Parts (`type: "subtask"`)
+- **Complete SwarmManager rewrite** (`swarm-manager.ts`)
+  - Now uses `client.session.promptAsync()` with `type: "subtask"` parts
+  - Each worker visible as separate subtask in chat
+  - True session isolation for each worker
+  - Named Result Capture with `as: "name"` and `$RESULT[name]`
+
+**Before vs After:**
+```typescript
+// BEFORE (not visible):
+await session.prompt({ sessionID, prompt, agent: "coder" })
+
+// AFTER (own subtask in chat):
+await client.session.promptAsync({
+  path: { id: sessionID },
+  body: {
+    parts: [{
+      type: "subtask",
+      agent: "coder",
+      model: modelSelection,
+      description: "Worker: Implement feature",
+      prompt: "...",
+      as: "worker_1"
+    }]
+  }
+})
+```
+
+#### Phase 6.2: Swarm Master Agent
+- **New Default Agent** (`agents/swarm-master.md`)
+  - Auto-detects when swarm is needed
+  - Simple tasks → Direct answer
+  - Complex tasks → Automatic `/swarm_run` call
+  - Intelligent task classification
+
+#### Phase 6.3: Complete Agent Army
+- **14 specialized agents** created:
+  - **Queens:** strategic-queen, tactical-queen, adaptive-queen
+  - **Workers:** researcher, coder, tester, analyst, architect, reviewer, optimizer, documenter
+  - **Swarm Specialists:** swarm-planner, swarm-implementer, swarm-reviewer
+  - **Master:** swarm-master (User Interface)
+- All consistently configured with .env models
+- Each agent with specific model, tools and instructions
+
+#### Phase 6.4: Subtask Parsing & Role Detection
+- **`parseSubtasksFromPlan()`** - Extracts numbered lists from planner output
+- **`detectRoleFromTask()`** - Auto-assigns roles based on task content
+- Dynamic worker assignment (no longer static)
+
+#### Phase 6.5: Comprehensive Logging System
+- **Logger** (`logger.ts`) with 5 levels: error, warn, info, debug, trace
+- Component-based loggers for detailed debugging
+- **`swarm_logs` Tool** - Retrieve and filter logs
+- Full transparency in subtask execution
+
+### Success Proof
+
+**Test Result:**
+```
+User: /swarm_run task="Analyze the swarm plugin architecture"
+
+Output:
+✅ Strategic-Queen creates plan (5 subtasks) - VISIBLE as subtask
+✅ Swarm-Implementer performs analysis - VISIBLE as subtask  
+✅ 764-line architecture report generated
+✅ All components documented
+```
+
+### Technical Details
+
+**Key Insight from subtask2:**
+- OpenCode's `task` Tool doesn't show subtasks
+- `type: "subtask"` Parts create true subtask sessions
+- This is undocumented but works perfectly
+
+**Implemented Patterns:**
+- Result Capture: `{as: "name"}` → `$RESULT[name]`
+- Parallel Execution: All workers simultaneously
+- Return Chains: Post-execution Actions
+- Progress Reporting: Live updates during execution
+
+#### New Files
+- `.opencode/swarm/core/logger.ts` (170 lines)
+- `.opencode/swarm/core/swarm-manager.ts` (completely rewritten, 450 lines)
+- `.opencode/plugins/swarm/tools/swarm-logs.ts` (54 lines)
+- All 14 agent definitions
+
+---
+
+## 0.5.0 - 2026-02-28
+
+### Phase 5: Advanced Features ✅
+
+#### Phase 5.1: MCP-Server Integration
+- **MCP Types** (`mcp.ts`) - MCP protocol types
+- **MCP Server** (`createMCPServer()`) - Resource, Tool & Prompt registration
+- **MCP Client** (`createMCPClient()`) - Resource access, Tool execution, Prompt retrieval
+- **Swarm Integration** (`createSwarmMCPAdapter()`) - Claude Desktop compatibility
+
+#### Phase 5.2: Dynamic Agent Discovery
+- **Service Registry** (`discovery.ts`) - Agent registration and discovery
+- **Load Balancing** - `selectLeastLoaded()`, `selectRoundRobin()`
+- **Health Monitoring** - Automatic health-checks, TTL-based cleanup
+- **Agent Watcher** (`createAgentWatcher()`) - Real-time updates
+- **Hot Reload** (`createHotReloadManager()`) - Automatic reload
+
+#### Phase 5.3: Swarm-Mail System
+- **Mailbox** (`mail.ts`) - Per-agent mailbox with priorities
+- **Mail Features** - Threading, Attachments, Metadata
+- **Mail Server** (`createMailServer()`) - Multi-agent routing
+- **Templates** (`createTemplate()`, `renderTemplate()`)
+
+#### Phase 5.4: Dashboard/API
+- **Dashboard API** (`dashboard.ts`) - Aggregated statistics
+- **Control Operations** - Pause/Resume/Cancel/Retry
+- **Export** - Prometheus/JSON export
+- **REST Server** (`createRESTServer()`)
+- **WebSocket Server** (`createWebSocketServer()`)
+- **HTML Dashboard** (`generateDashboardHTML()`)
+
+---
+
+## 0.4.6 - 2026-02-28
+
+### Phase 4: Production-Readiness ✅
+
+#### Phase 4.1: Extended Observability
+- **Metrics Collector** (`metrics.ts`) - Prometheus-compatible export
+- **Tracer** (`tracer()`) - Distributed tracing with parent-child spans
+
+#### Phase 4.2: SQLite Task Queue (In-Memory)
+- **Task Queue** (`queue.ts`) - Priority-based scheduling
+- **Queue Processor** (`createQueueProcessor()`) - Concurrency limiting
+
+#### Phase 4.3: State Recovery & Checkpointing
+- **State Manager** (`recovery.ts`) - JSON persistence
+- **Recovery Coordinator** - Crash recovery
+
+#### Phase 4.4: Integration Tests
+- **E2E Test Suite** (`integration.spec.ts`) - 689 lines
+
+#### Phase 4.5: Error Enrichment
+- **Hierarchical Error Classes** (`errors.ts`)
+- **Error Codes** - 25+ specific codes
+- **Recovery Suggestions** - Automatic suggestions
+- **Error Factory** - `ErrorFactory.sessionNotFound()`, etc.
+- **Retry Decorator** - `withRetry()` with exponential backoff
+
+---
+
+## 0.4.5 - 2026-02-28
+
+### Phase 3: True Subagent Implementation ✅
+
+#### Phase 3.1: Task/Subagent Dispatcher
+- **Dispatcher** (`dispatcher.ts`) with 3 strategies
+- Native → Fork → Prompt graceful degradation
+
+#### Phase 3.2: Worker Isolation
+- **Worker Factory** (`worker.ts`) - Isolated worker creation
+- **Worker Pool** - Queue and concurrency limiting
+
+#### Phase 3.3: Message Bus
+- **Message Bus** (`messaging.ts`) - P2P, Broadcast, Request/Response
+
+#### Phase 3.4: Context Propagation
+- **Context Manager** (`context.ts`) - Parent-child hierarchy
+- **Checkpoint Manager** - Automatic checkpoints
+
+---
+
+## 0.4.4 - 2026-02-28
+
+### Phase 2.5: Workflow Engine (subtask2 concepts) ✅
+
+- **Complete Workflow Engine** - Sequential, Parallel, Hierarchical
+- **Result Store** (`results.ts`) - `$RESULT[name]` variable resolution
+- **Workflow Executor** (`executor.ts`) - Step execution with retry
+- **Loop Engine** (`loop.ts`) - LLM-based loop evaluation
+- **Parallel Coordinator** (`parallel.ts`) - Dependency graph, topological sort
+
+---
+
+## 0.4.3 - 2026-02-28
+
+### Phase 2.4: Unified Adapter Interface ✅
+
+- **Unified Adapter Interface** (`base.ts`)
+- `BaseSessionAdapter` interface with required/optional methods
+- `AdapterError` class with error codes
+- Retry logic with exponential backoff
+
+---
+
+## 0.4.2 - 2026-02-28
+
+### Phase 2.3: Agent Registry as Separate Module ✅
+
+- **Agent Registry** (`registry.ts`) - CRUD with capability detection
+- New tools: `swarm_agent_list`, `swarm_agent_find`, `swarm_agent_get`, `swarm_agent_stats`
+
+---
+
+## 0.4.1 - 2026-02-28
+
+### Phase 2.2: Config System with Zod Validation ✅
+
+- **Extended Config System** - File support and hot-reload
+- `createConfigManager()` with `.opencode/swarm.config.json`
+- Env validation with clear error messages
+
+---
+
+## 0.4.0 - 2026-02-28
+
+### Phase 2.1: Plugin Modularization ✅
+
+- **Complete Architecture Overhaul** (950 → ~580 lines, 12 modules)
+- New structure: `swarm/{types,constants,config,utils,tools,hooks}/`
+
+---
+
+## 0.3.1 - 2026-02-28
+
+### Hotfix
+
+- **Fixed ProviderModelNotFoundError**
+- Models only set when defined
+
+---
+
+## 0.3.0 - 2026-02-28
+
+### Phase 1: Critical Stability Fixes ✅
+
+- **Fixed binding bug** (`this._client is undefined`)
+- **Concurrency Limiter** (max 3 parallel workers)
+- **Retry Logic** with exponential backoff
+
+---
+
+## 0.2.0 - 2026-02-28
+
+- Coordinator guard enforcement with documented hooks
+- Compaction-resume context injection
+- Diagnostic tools: `swarm_trace`, `swarm_guard_status`
+- Child-session isolation for `swarm_run`
+- Prompt timeout handling
+- Role-based model mapping via env
+
+---
+
+## 0.1.0 - 2026-02-27
+
+- Initial production-focused Swarm plugin baseline
+- SDK-first session orchestration
+- Security hooks for tool execution
+- Persistent JSON memory
+- Session utility tools
+
+</details>
+
+---
+
+<p align="center">
+  <a href="README.md">← Back to README</a>
+</p>

package/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 Kevin (RealBlxckCodex)
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

package/README.md

@@ -0,0 +1,504 @@
+# 🐝 OpenCode Swarm Plugin
+
+> **Multi-Agent Orchestration System** with true subagent support for OpenCode AI
+
+[![Version](https://img.shields.io/badge/version-0.6.0-blue.svg)](CHANGELOG.md)
+[![Status](https://img.shields.io/badge/status-production--ready-green.svg)](ROADMAP.md)
+[![License](https://img.shields.io/badge/license-MIT-yellow.svg)](LICENSE)
+
+---
+
+<details>
+<summary>🇩🇪 <strong>Deutsch</strong> | German</summary>
+
+## 🎯 Was ist OpenCode Swarm?
+
+OpenCode Swarm ist ein **produktionsreifes Multi-Agent-Orchestrierungssystem**, das komplexe Entwicklungsaufgaben durch koordinierte Teams von KI-Agenten ausführt. Im Gegensatz zu einfachen Agent-Ketten implementiert Swarm ein **echtes Subagent-Pattern**, bei dem jeder Worker als isolierter Subtask mit voller Sichtbarkeit in der OpenCode-Chat-Oberfläche läuft.
+
+### ✨ Hauptfunktionen
+
+- 🎭 **14 Spezialisierte Agents** - Von strategischen Planern bis zu Code-Reviewern
+- 🔄 **Echte Subagent-Orchestrierung** - Jeder Worker als separater Subtask sichtbar
+- 🏗️ **Hierarchische & Mesh-Workflows** - Flexible Ausführungsmuster
+- 🗳️ **Konsens-Mechanismen** - Byzantinisch, Abstimmung und gewichteter Konsens
+- 🧠 **Intelligente Modellauswahl** - Automatische Model-Tier-Zuweisung
+- 📊 **Kosten-Tracking** - API-Nutzung pro Run überwachen
+- 🔒 **Security Guard** - Eingabe-Sanitisierung und Validierung
+- 📈 **Umfassendes Logging** - Debuggen mit detaillierten Ausführungs-Logs
+
+## 🚀 Schnellstart
+
+### Installation
+
+```bash
+# Repository klonen
+git clone https://github.com/RealBlxckCodex/OpenSwarm.git
+cd OpenSwarm
+
+# Abhängigkeiten installieren
+npm install
+
+# Umgebung konfigurieren
+cp .env.example .env
+# .env mit deinen API-Keys bearbeiten
+```
+
+### Konfiguration
+
+Erstelle `.env` mit deinen bevorzugten Modellen:
+
+```env
+# Model Tiers
+THINKING=kimi-k2-thinking
+CODING=kimi-k2-5
+STRONG=gpt-5-codex
+REVIEWER=gpt-5-codex
+```
+
+### Nutzung
+
+```bash
+# Swarm-Run starten
+/swarm_run task="Benutzer-Authentifizierungssystem implementieren"
+
+# Ausführungs-Logs prüfen
+/swarm_logs
+
+# Mit spezifischem Modus ausführen
+/swarm_run task="Datenbank-Layer refactoren" mode=hierarchical
+```
+
+## 🏗️ Architektur
+
+### Drei-Ebenen-Agent-System
+
+```
+┌─────────────────────────────────────────┐
+│         Benutzer-Oberflächen-Ebene      │
+│         (swarm-master)                  │
+└─────────────────┬───────────────────────┘
+                  │
+┌─────────────────▼───────────────────────┐
+│         Strategische Ebene              │
+│  ┌─────────┐ ┌─────────┐ ┌─────────┐   │
+│  │strategic│ │tactical │ │adaptive │   │
+│  │ -queen  │ │ -queen  │ │ -queen  │   │
+│  └────┬────┘ └────┬────┘ └────┬────┘   │
+└───────┼──────────┼──────────┼────────┘
+        │          │          │
+┌───────▼──────────▼──────────▼────────┐
+│         Ausführungs-Ebene             │
+│  researcher │ coder │ tester │        │
+│  architect  │reviewer│optimizer│       │
+│  analyst    │documenter              │
+└───────────────────────────────────────┘
+```
+
+</details>
+
+---
+
+<details open>
+<summary>🇬🇧 <strong>English</strong> | Englisch</summary>
+
+## 🎯 What is OpenCode Swarm?
+
+OpenCode Swarm is a **production-ready multi-agent orchestration system** that enables complex development tasks to be executed by coordinated teams of AI agents. Unlike simple agent chains, Swarm implements a **true subagent pattern** where each worker runs as an isolated subtask with full visibility in the OpenCode chat interface.
+
+### ✨ Key Features
+
+- 🎭 **14 Specialized Agents** - From strategic planners to code reviewers
+- 🔄 **True Subagent Orchestration** - Each worker visible as separate subtask
+- 🏗️ **Hierarchical & Mesh Workflows** - Flexible execution patterns
+- 🗳️ **Consensus Mechanisms** - Byzantine, voting, and weighted consensus
+- 🧠 **Smart Model Selection** - Automatic model tier assignment
+- 📊 **Cost Tracking** - Monitor API usage per run
+- 🔒 **Security Guard** - Input sanitization and validation
+- 📈 **Comprehensive Logging** - Debug with detailed execution logs
+
+## 🚀 Quick Start
+
+### Installation
+
+```bash
+# Clone the repository
+git clone https://github.com/RealBlxckCodex/OpenSwarm.git
+cd OpenSwarm
+
+# Install dependencies
+npm install
+
+# Configure environment
+cp .env.example .env
+# Edit .env with your API keys
+```
+
+### Configuration
+
+Create `.env` with your preferred models:
+
+```env
+# Model Tiers
+THINKING=kimi-k2-thinking
+CODING=kimi-k2-5
+STRONG=gpt-5-codex
+REVIEWER=gpt-5-codex
+```
+
+### Usage
+
+```bash
+# Start a swarm run
+/swarm_run task="Implement user authentication system"
+
+# Check execution logs
+/swarm_logs
+
+# Run with specific mode
+/swarm_run task="Refactor database layer" mode=hierarchical
+```
+
+## 🏗️ Architecture
+
+### Three-Tier Agent System
+
+```
+┌─────────────────────────────────────────┐
+│         User Interface Layer            │
+│         (swarm-master)                  │
+└─────────────────┬───────────────────────┘
+                  │
+┌─────────────────▼───────────────────────┐
+│         Strategic Layer                 │
+│  ┌─────────┐ ┌─────────┐ ┌─────────┐   │
+│  │strategic│ │tactical │ │adaptive │   │
+│  │ -queen  │ │ -queen  │ │ -queen  │   │
+│  └────┬────┘ └────┬────┘ └────┬────┘   │
+└───────┼──────────┼──────────┼────────┘
+        │          │          │
+┌───────▼──────────▼──────────▼────────┐
+│         Execution Layer               │
+│  researcher │ coder │ tester │        │
+│  architect  │reviewer│optimizer│       │
+│  analyst    │documenter              │
+└───────────────────────────────────────┘
+```
+
+</details>
+
+---
+
+<details>
+<summary>🇩🇪 <strong>Workflow-Modi</strong> | Workflow Modes</summary>
+
+**Hierarchischer Modus:**
+```
+strategic-queen → planner → workers → consensus → reviewer
+```
+
+**Mesh-Modus:**
+```
+Alle Worker führen parallel mit Peer-Konsens aus
+```
+
+</details>
+
+<details open>
+<summary>🇬🇧 <strong>Workflow Modes</strong> | Workflow-Modi</summary>
+
+**Hierarchical Mode:**
+```
+strategic-queen → planner → workers → consensus → reviewer
+```
+
+**Mesh Mode:**
+```
+All workers execute in parallel with peer consensus
+```
+
+</details>
+
+---
+
+<details>
+<summary>🇩🇪 <strong>Verfügbare Agents</strong> | Available Agents</summary>
+
+### Queens (Strategische Ebene)
+- **strategic-queen** - Hochrangige Planung & Dekomposition
+- **tactical-queen** - Ausführungs-fokussierte Planung
+- **adaptive-queen** - Selbst-optimierend mit Feedback-Loops
+
+### Worker (Ausführungs-Ebene)
+- **researcher** - Codebase-Analyse & Dokumentation
+- **coder** - Implementierungs-Spezialist
+- **tester** - Test-Generierung & Validierung
+- **architect** - System-Design & Patterns
+- **reviewer** - Code-Review & Qualitätssicherung
+- **analyst** - Anforderungsanalyse
+- **optimizer** - Performance-Optimierung
+- **documenter** - Dokumentations-Generierung
+
+### Swarm-Spezialisten
+- **swarm-planner** - Swarm-spezifische Planung
+- **swarm-implementer** - Multi-Agent-Ausführung
+- **swarm-reviewer** - Swarm-Output-Review
+- **swarm-master** - Benutzeroberfläche & Delegation
+
+</details>
+
+<details open>
+<summary>🇬🇧 <strong>Available Agents</strong> | Verfügbare Agents</summary>
+
+### Queens (Strategic Layer)
+- **strategic-queen** - High-level planning & decomposition
+- **tactical-queen** - Execution-focused planning
+- **adaptive-queen** - Self-optimizing with feedback loops
+
+### Workers (Execution Layer)
+- **researcher** - Codebase analysis & documentation
+- **coder** - Implementation specialist
+- **tester** - Test generation & validation
+- **architect** - System design & patterns
+- **reviewer** - Code review & quality assurance
+- **analyst** - Requirements analysis
+- **optimizer** - Performance optimization
+- **documenter** - Documentation generation
+
+### Swarm Specialists
+- **swarm-planner** - Swarm-specific planning
+- **swarm-implementer** - Multi-agent execution
+- **swarm-reviewer** - Swarm output review
+- **swarm-master** - User interface & delegation
+
+</details>
+
+---
+
+<details>
+<summary>🇩🇪 <strong>Befehle</strong> | Commands</summary>
+
+| Befehl | Beschreibung |
+|--------|-------------|
+| `/swarm_run` | Task mit dem Swarm ausführen |
+| `/swarm_logs` | Ausführungs-Logs abrufen |
+| `/swarm_debug` | Diagnose ausführen |
+
+### Befehls-Optionen
+
+```bash
+/swarm_run 
+  task="Deine Task-Beschreibung"    # Erforderlich
+  mode=hierarchical|mesh            # Workflow-Modus
+  consensus=byzantine|voting        # Konsens-Typ
+  depth=3                           # Planungs-Tiefe
+  verbose=true                      # Detaillierte Ausgabe
+```
+
+</details>
+
+<details open>
+<summary>🇬🇧 <strong>Commands</strong> | Befehle</summary>
+
+| Command | Description |
+|---------|-------------|
+| `/swarm_run` | Execute a task with the swarm |
+| `/swarm_logs` | Retrieve execution logs |
+| `/swarm_debug` | Run diagnostics |
+
+### Command Options
+
+```bash
+/swarm_run 
+  task="Your task description"    # Required
+  mode=hierarchical|mesh          # Workflow mode
+  consensus=byzantine|voting      # Consensus type
+  depth=3                         # Planning depth
+  verbose=true                    # Detailed output
+```
+
+</details>
+
+---
+
+<details>
+<summary>🇩🇪 <strong>Projektstruktur</strong> | Project Structure</summary>
+
+```
+OpenCode-Swarm/
+├── .opencode/
+│   ├── agents/              # 14 Agent-Definitionen
+│   ├── plugins/swarm/       # Plugin-Implementierung
+│   ├── swarm/
+│   │   ├── core/            # Core-Orchestrierung (~11.200 Zeilen)
+│   │   ├── adapters/        # Plattform-Adapter
+│   │   └── modules/         # Erweiterungssystem
+│   └── commands/            # Debug-Befehle
+├── .github/                 # GitHub-Templates
+├── CHANGELOG.md            # Versions-Historie
+├── ROADMAP.md              # Feature-Roadmap
+└── FEATURE_BRAINSTORMING.md # Zukünftige Features
+```
+
+</details>
+
+<details open>
+<summary>🇬🇧 <strong>Project Structure</strong> | Projektstruktur</summary>
+
+```
+OpenCode-Swarm/
+├── .opencode/
+│   ├── agents/              # 14 agent definitions
+│   ├── plugins/swarm/       # Plugin implementation
+│   ├── swarm/
+│   │   ├── core/            # Core orchestration (~11,200 lines)
+│   │   ├── adapters/        # Platform adapters
+│   │   └── modules/         # Extension system
+│   └── commands/            # Debug commands
+├── .github/                 # GitHub templates
+├── CHANGELOG.md            # Version history
+├── ROADMAP.md              # Feature roadmap
+└── FEATURE_BRAINSTORMING.md # Future ideas
+```
+
+</details>
+
+---
+
+<details>
+<summary>🇩🇪 <strong>Dokumentation</strong> | Documentation</summary>
+
+| Dokument | Beschreibung |
+|----------|-------------|
+| [ROADMAP.md](ROADMAP.md) | Entwicklungs-Roadmap & Meilensteine |
+| [CHANGELOG.md](CHANGELOG.md) | Versions-Historie & Änderungen |
+| [FEATURE_BRAINSTORMING.md](FEATURE_BRAINSTORMING.md) | Zukünftige Feature-Ideen |
+| [SWARM_ARCHITECTURE_REPORT.md](SWARM_ARCHITECTURE_REPORT.md) | Detaillierte Architektur-Analyse |
+| [TEST_PLAN.md](TEST_PLAN.md) | Test-Strategie & Validierung |
+
+</details>
+
+<details open>
+<summary>🇬🇧 <strong>Documentation</strong> | Dokumentation</summary>
+
+| Document | Description |
+|----------|-------------|
+| [ROADMAP.md](ROADMAP.md) | Development roadmap & milestones |
+| [CHANGELOG.md](CHANGELOG.md) | Version history & changes |
+| [FEATURE_BRAINSTORMING.md](FEATURE_BRAINSTORMING.md) | Future feature ideas |
+| [SWARM_ARCHITECTURE_REPORT.md](SWARM_ARCHITECTURE_REPORT.md) | Detailed architecture analysis |
+| [TEST_PLAN.md](TEST_PLAN.md) | Testing strategy & validation |
+
+</details>
+
+---
+
+<details>
+<summary>🇩🇪 <strong>Tests</strong> | Testing</summary>
+
+```bash
+# Einfachen Test ausführen
+/swarm_run task="Teste Grundfunktionalität"
+
+# Parallelen Ausführungs-Test ausführen
+/swarm_run task="Teste parallele Worker" mode=mesh
+
+# System-Status prüfen
+/swarm_debug
+```
+
+</details>
+
+<details open>
+<summary>🇬🇧 <strong>Testing</strong> | Tests</summary>
+
+```bash
+# Run simple test
+/swarm_run task="Test basic functionality"
+
+# Run parallel execution test
+/swarm_run task="Test parallel workers" mode=mesh
+
+# Check system health
+/swarm_debug
+```
+
+</details>
+
+---
+
+<details>
+<summary>🇩🇪 <strong>Performance</strong> | Performance</summary>
+
+- **Ausführungszeit:** 4-5 Minuten für komplexe Tasks
+- **Erfolgsrate:** >90%% für gut definierte Tasks
+- **Kosteneffizienz:** Intelligente Model-Tier-Auswahl
+- **Parallelisierung:** Bis zu 8 gleichzeitige Worker
+
+</details>
+
+<details open>
+<summary>🇬🇧 <strong>Performance</strong> | Performance</summary>
+
+- **Execution Time:** 4-5 minutes for complex tasks
+- **Success Rate:** >90% for well-defined tasks
+- **Cost Efficiency:** Smart model tier selection
+- **Parallelism:** Up to 8 concurrent workers
+
+</details>
+
+---
+
+<details>
+<summary>🇩🇪 <strong>Mitwirken</strong> | Contributing</summary>
+
+Mitwirkungen sind willkommen! Siehe unsere [GitHub Issues](https://github.com/RealBlxckCodex/OpenSwarm/issues) für:
+
+- 🐛 Bug-Reports
+- 💡 Feature-Requests
+- 📖 Dokumentations-Verbesserungen
+
+## 📝 Lizenz
+
+Dieses Projekt ist unter der MIT-Lizenz lizenziert.
+
+## 🙏 Danksagungen
+
+- Gebaut für [OpenCode AI](https://opencode.ai)
+- Inspiriert durch Multi-Agent-Forschung
+- Dank an das OpenCode-Team für das `type: "subtask"` Pattern
+
+</details>
+
+<details open>
+<summary>🇬🇧 <strong>Contributing</strong> | Mitwirken</summary>
+
+Contributions are welcome! Please see our [GitHub Issues](https://github.com/RealBlxckCodex/OpenSwarm/issues) for:
+
+- 🐛 Bug reports
+- 💡 Feature requests
+- 📖 Documentation improvements
+
+## 📝 License
+
+This project is licensed under the MIT License.
+
+## 🙏 Acknowledgments
+
+- Built for [OpenCode AI](https://opencode.ai)
+- Inspired by multi-agent research
+- Thanks to the OpenCode team for the `type: "subtask"` pattern
+
+</details>
+
+---
+
+<p align="center">
+  <strong>🐝 Swarm Intelligence for Better Code 🐝</strong><br>
+  <em>🐝 Schwarm-Intelligenz für besseren Code 🐝</em>
+</p>
+
+<p align="center">
+  <a href="CHANGELOG.md">Changelog</a> •
+  <a href="ROADMAP.md">Roadmap</a> •
+  <a href="FEATURE_BRAINSTORMING.md">Ideas/Ideen</a>
+</p>

package/package.json

@@ -0,0 +1,58 @@
+{
+  "name": "@realblxckcodex/opencode-swarm",
+  "version": "0.6.0",
+  "description": "Multi-Agent Orchestration System with true subagent support for OpenCode AI",
+  "type": "module",
+  "main": "./dist/index.js",
+  "types": "./dist/index.d.ts",
+  "files": [
+    "dist/**/*",
+    "README.md",
+    "LICENSE",
+    "CHANGELOG.md"
+  ],
+  "scripts": {
+    "build": "tsc -p tsconfig.json",
+    "test": "tsx --test .opencode/swarm/core/swarm-core.spec.ts",
+    "typecheck": "tsc -p tsconfig.json --noEmit",
+    "lint": "tsc --noEmit",
+    "release:check": "npm run typecheck && npm run test",
+    "ci:local": "npm run release:check",
+    "prepublishOnly": "npm run build && npm run release:check"
+  },
+  "keywords": [
+    "opencode",
+    "swarm",
+    "multi-agent",
+    "orchestration",
+    "ai",
+    "llm",
+    "typescript",
+    "plugin"
+  ],
+  "author": "Kevin (RealBlxckCodex) <your.email@example.com>",
+  "license": "MIT",
+  "repository": {
+    "type": "git",
+    "url": "git+https://github.com/RealBlxckCodex/OpenSwarm.git"
+  },
+  "bugs": {
+    "url": "https://github.com/RealBlxckCodex/OpenSwarm/issues"
+  },
+  "homepage": "https://github.com/RealBlxckCodex/OpenSwarm#readme",
+  "engines": {
+    "node": ">=18.0.0"
+  },
+  "devDependencies": {
+    "@opencode-ai/plugin": "latest",
+    "@types/node": "^24.0.0",
+    "tsx": "^4.19.0",
+    "typescript": "^5.8.0"
+  },
+  "dependencies": {
+    "zod": "^4.3.6"
+  },
+  "publishConfig": {
+    "access": "public"
+  }
+}