opencode-orchestrator 1.0.10 → 1.0.12

package/README.md

@@ -12,12 +12,6 @@
 
 ---
 
-### 🌌 **The End of Sequential Limits. The Era of Fractal Mastery.**
-
-Where traditional agents hit a sequential wall, we deploy a fractal swarm. OpenCode Orchestrator is the world's most advanced autonomous engineering platform—a **Titan-Class Execution Engine** designed to conquer missions that break conventional AI. 
-
-By fusing **Fractal Swarm Intelligence (HPFA™)** with **Continuous Multi-Stage Verification Pipeline (MSVP™)**, it delivers a level of scale, velocity, and architectural absolute previously considered impossible. Welcome to the final form of autonomous development. No bottlenecks. No limits. Only pure execution.
-
 ## ⚡ Quick Start
 
 ```bash
@@ -26,9 +20,16 @@ npm install -g opencode-orchestrator
 
 Then in OpenCode, launch your mission:
 ```bash
-/extreme-mission "Architect and Build a Diablo-like Web Engine"
+/swarm "Architect and Build a Diablo-like Web Game"
 ```
 
+### 🌌 **The End of Sequential Limits. The Era of Fractal Mastery.**
+
+Where traditional agents hit a sequential wall, we deploy a fractal swarm. OpenCode Orchestrator is the world's most advanced autonomous engineering platform—a **Titan-Class Execution Engine** designed to conquer missions that break conventional AI. 
+
+By fusing **Fractal Swarm Intelligence (HPFA™)** with **Continuous Multi-Stage Verification Pipeline (MSVP™)**, it delivers a level of scale, velocity, and architectural absolute previously considered impossible. Welcome to the final form of autonomous development. No bottlenecks. No limits. Only pure execution.
+
+
 ## 💎 The Four Pillars of Excellence
 
 We’ve combined industrial-grade reliability with cutting-edge parallel intelligence to redefine what’s possible in AI coding.
@@ -74,20 +75,20 @@ We’ve combined industrial-grade reliability with cutting-edge parallel intelli
 
 ## 🏛️ Hyper-Parallel Architecture: HPFA & MSVP
 
-The orchestrator eliminates the linear bottleneck of AI development through two breakthrough technologies:
+The orchestrator eliminates the linear reasoning bottleneck of conventional AI agents by shifting from sequential execution to a **Distributed Runtime Unit** paradigm.
 
-### 🧬 **HPFA (Hyper-Parallel Fractal Architecture)**
-Imagine an agent that doesn't just work, but **thinks like a Hive Mind**. 
-*   **Fractal Delegation**: Workers are no longer just "doers." They are mini-planners that can recursively spawn their own sub-agents to handle complexity at every level.
-*   **Speculative Racing**: Ambiguous tasks are solved using multiple strategies in parallel. The fastest, most accurate path wins, while others are pruned—minimizing latency and maximizing quality.
+### 🧬 **HPFA™ (Hyper-Parallel Fractal Architecture)**
+HPFA defines a recursive problem-solving model that decomposes massive codebases into atomic task units through a fractal hierarchy.
+*   **Fractal Delegation**: Workers are instantiated as autonomous controllers with localized `TaskLauncher` contexts. This enables recursive spawning of sub-agents up to 8 levels deep, mapping complex architectural hierarchies into a parallel execution grid.
+*   **Cognitive Concurrency Control**: A dedicated state-machine manages up to 50 concurrent sessions. It monitors success/failure trajectories to dynamically scale concurrency slots, ensuring maximum throughput while maintaining strict rate-limit compliance.
 
-### 🛡️ **MSVP (Multi-Stage Verification Pipeline)**
-Speed is nothing without accuracy. MSVP ensures your code is production-ready at every second.
-*   **Shadow Parallel Review**: As soon as a file is written, a parallel reviewer session is already validating it before the main task even finishes.
-*   **Barrier-Sync Integration**: A 4-stage verification gate that prevents "architectural drift" by forcing a global synchronization barrier before final sealing.
+### 🛡️ **MSVP™ (Multi-Stage Verification Pipeline)**
+MSVP ensures continuous integration integrity by decoupling verification from the implementation lifecycle, treating code review as a parallel background-thread.
+*   **Shadow Review Pipeline**: Upon every file-write operation, a parallel reviewer session is instantly snapshotted (Unit-level Verification). This zero-latency audit identifies defects the millisecond they are introduced, without blocking the primary implementation thread.
+*   **Global Barrier-Sync Protocol**: Implements a synchronized verification gate. The system enforces a blocking "Barrier" that requires consistency across all distributed units before transitioning to the final Master Review (E2E Integration) and Mission Sealing.
 
 ```
-              /extreme-mission "Build Complex System"
+              /swarm "Build Diablo Game"
                            │
            ╔═══════════════════════════════════════╗
            ║     🎯 COMMANDER — Orchestrator       ║
@@ -148,14 +149,14 @@ Speed is nothing without accuracy. MSVP ensures your code is production-ready at
 
 | Feature | What It Does |
 |:---------|:-------------|
-| 🚀 **MSVP Pipeline** | Parallel unit reviews triggered immediately after worker tasks |
-| 🧬 **Fractal Spawning** | Workers can recursively spawn sub-missions for complex tasks |
-| ⚡ **20+ Parallel Tasks** | High-intensity execution with safety concurrency limits |
-| 🔄 **Non-Stop Recovery** | WAL-based persistence (resumes tasks after crashes) |
-| 🔥 **Multi-File Ops** | Work on different files at the same time |
-| 🛡️ **Self-Scaling** | Dynamic concurrency limits based on success/failure |
-| 🩹 **Memory Integrity** | Strict resource cleanup prevents leaks in long sessions |
-| 🏗️ **Architectural Modularity**| Enforces language-agnostic clean code principles |
+| 🚀 **MSVP™ Pipeline** | Zero-latency unit reviews triggered the millisecond code is written |
+| 🧬 **Fractal Spawning** | Workers recursively spawn sub-agents for infinite structural depth |
+| ⚡ **50 Parallel Tasks** | Titan-class density—run massive swarms without breaking a sweat |
+| 🔄 **Non-Stop Recovery** | WAL-powered cognitive persistence; resumes instantly after any crash |
+| 🔥 **Multi-File Ops** | Massive parallel edits with cross-file context integrity |
+| 🛡️ **Self-Scaling** | Dynamic AI-driven concurrency management based on success velocity |
+| 🩹 **Memory Integrity** | 100% reclamation; engineered for sessions with 10k+ iterations |
+| 🏗️ **Clean-Code Engine**| Enforces industrial modularity and clean architectural patterns |
 
 ---
 

package/dist/index.js

@@ -288,11 +288,11 @@ var PARALLEL_TASK = {
   // Task lifecycle (24 hours for long tasks)
   TTL_MS: 24 * TIME.HOUR,
   CLEANUP_DELAY_MS: 10 * TIME.MINUTE,
-  MAX_DEPTH: 5,
-  // Increased for fractal recursion
+  MAX_DEPTH: 8,
+  // Titan-Class depth for complex fractal missions
   // Concurrency limits (Aggressive for intense processing)
-  DEFAULT_CONCURRENCY: 5,
-  MAX_CONCURRENCY: 20,
+  DEFAULT_CONCURRENCY: 10,
+  MAX_CONCURRENCY: 50,
   // Sync polling (for delegate_task sync mode)
   // Optimized: Reduced polling frequency while relying more on events
   SYNC_TIMEOUT_MS: 5 * TIME.MINUTE,
@@ -631,7 +631,7 @@ var PROMPTS = {
 
 // src/shared/message/constants/slash-commands.ts
 var COMMAND_NAMES = {
-  TASK: "extreme-mission",
+  TASK: "swarm",
   PLAN: "plan",
   STATUS: "status",
   STOP: "stop",
@@ -15532,7 +15532,7 @@ $ARGUMENTS
 </execution_rules>
 </mission>`;
 var COMMANDS = {
-  "extreme-mission": {
+  "swarm": {
     description: "MISSION MODE - Execute task autonomously until complete",
     template: MISSION_MODE_TEMPLATE,
     argumentHint: '"mission goal"'
@@ -15593,7 +15593,7 @@ THINK \u2192 PLAN \u2192 DELEGATE \u2192 EXECUTE \u2192 VERIFY \u2192 COMPLETE
 
 ## Usage
 - Select **${AGENT_NAMES.COMMANDER}** and type your request
-- Or use \`/extreme-mission "your mission"\` explicitly
+- Or use \`/swarm "your mission"\` explicitly
 - ${AGENT_NAMES.COMMANDER} automatically coordinates all agents`
   }
 };
@@ -16249,7 +16249,7 @@ var ConcurrencyController = class {
       this.failureCount.set(key, 0);
       if (streak % 5 === 0) {
         const currentLimit = this.getConcurrencyLimit(key);
-        if (currentLimit < 20) {
+        if (currentLimit < PARALLEL_TASK.MAX_CONCURRENCY) {
           this.setLimit(key, currentLimit + 1);
           log(`[concurrency] Auto-scaling UP for ${key}: ${currentLimit + 1}`);
         }
@@ -17060,6 +17060,11 @@ var TaskLauncher = class {
    * Prepare task: Create session and registration without blocking on concurrency
    */
   async prepareTask(input) {
+    const currentDepth = input.depth ?? 0;
+    if (currentDepth >= PARALLEL_TASK.MAX_DEPTH) {
+      log(`[task-launcher.ts] Task depth limit reached (${currentDepth}/${PARALLEL_TASK.MAX_DEPTH}). Generation blocked.`);
+      throw new Error(`Maximum task depth (${PARALLEL_TASK.MAX_DEPTH}) reached. To prevent infinite recursion, no further sub-tasks can be spawned.`);
+    }
     const createResult = await this.client.session.create({
       body: {
         parentID: input.parentSessionID,
@@ -17908,7 +17913,10 @@ var createDelegateTaskTool = (manager, client) => tool({
   async execute(args, context) {
     const { agent, description, prompt, background, resume, mode, groupID } = args;
     const ctx = context;
-    log(`${PARALLEL_LOG.DELEGATE_TASK} execute() called`, { agent, description, background, resume, mode, groupID, parentSession: ctx.sessionID });
+    const allTasks = manager.getAllTasks();
+    const parentTask = allTasks.find((t) => t.sessionID === ctx.sessionID);
+    const parentDepth = parentTask?.depth ?? 0;
+    log(`${PARALLEL_LOG.DELEGATE_TASK} execute() called`, { agent, description, background, resume, mode, groupID, parentSession: ctx.sessionID, depth: parentDepth });
     const sessionClient = client;
     if (background === void 0) {
       return `${OUTPUT_LABEL.ERROR} 'background' parameter is REQUIRED.`;
@@ -17924,7 +17932,9 @@ var createDelegateTaskTool = (manager, client) => tool({
           description,
           // Assuming description is needed for resume context
           mode,
-          groupID
+          groupID,
+          depth: parentDepth
+          // Preserve depth on resume
         };
         const launchResult = await manager.launch(input);
         const task = Array.isArray(launchResult) ? launchResult[0] : launchResult;
@@ -17964,7 +17974,8 @@ ${text || "(No output)"}`;
           prompt,
           parentSessionID: ctx.sessionID,
           mode,
-          groupID
+          groupID,
+          depth: parentDepth
         });
         const task = Array.isArray(launchResult) ? launchResult[0] : launchResult;
         presets.taskStarted(task.id, agent);
@@ -20538,7 +20549,7 @@ ${claudeRules}`;
 function detectSlashCommand(text) {
   const match = text.trim().match(/^\/([a-zA-Z0-9_-]+)(?:\s+(.*))?$/);
   if (!match) return null;
-  return { command: match[1], args: match[2] || "" };
+  return { command: match[1].toLowerCase(), args: match[2] || "" };
 }
 
 // src/utils/formatting/timestamp.ts
@@ -20576,7 +20587,7 @@ function createChatMessageHandler(ctx) {
     if (sessionID) {
       handleUserMessage(sessionID);
     }
-    if (agentName === AGENT_NAMES.COMMANDER) {
+    if (agentName === AGENT_NAMES.COMMANDER.toLowerCase()) {
       if (!sessions.has(sessionID)) {
         const now = Date.now();
         sessions.set(sessionID, {
@@ -20610,7 +20621,7 @@ function createChatMessageHandler(ctx) {
     }
     if (parsed) {
       const command = COMMANDS[parsed.command];
-      if (command && agentName !== AGENT_NAMES.COMMANDER) {
+      if (command && agentName !== AGENT_NAMES.COMMANDER.toLowerCase()) {
         parts[textPartIndex].text = command.template.replace(
           /\$ARGUMENTS/g,
           parsed.args || PROMPTS.CONTINUE

package/dist/shared/message/constants/slash-commands.d.ts

@@ -2,7 +2,7 @@
  * Slash Commands (with slash prefix)
  */
 export declare const SLASH_COMMANDS: {
-    readonly TASK: "/extreme-mission";
+    readonly TASK: "/swarm";
     readonly PLAN: "/plan";
     readonly STATUS: "/status";
     readonly STOP: "/stop";
@@ -12,7 +12,7 @@ export declare const SLASH_COMMANDS: {
  * Command Names (without slash prefix, for comparison after parsing)
  */
 export declare const COMMAND_NAMES: {
-    readonly TASK: "extreme-mission";
+    readonly TASK: "swarm";
     readonly PLAN: "plan";
     readonly STATUS: "status";
     readonly STOP: "stop";

package/dist/shared/task/constants/parallel-task.d.ts

@@ -9,9 +9,9 @@ export declare const TASK_MODE: {
 export declare const PARALLEL_TASK: {
     readonly TTL_MS: number;
     readonly CLEANUP_DELAY_MS: number;
-    readonly MAX_DEPTH: 5;
-    readonly DEFAULT_CONCURRENCY: 5;
-    readonly MAX_CONCURRENCY: 20;
+    readonly MAX_DEPTH: 8;
+    readonly DEFAULT_CONCURRENCY: 10;
+    readonly MAX_CONCURRENCY: 50;
     readonly SYNC_TIMEOUT_MS: number;
     readonly POLL_INTERVAL_MS: 2000;
     readonly MIN_IDLE_TIME_MS: number;

package/package.json

@@ -2,7 +2,7 @@
   "name": "opencode-orchestrator",
   "displayName": "OpenCode Orchestrator",
   "description": "Distributed Cognitive Architecture for OpenCode. Turns simple prompts into specialized multi-agent workflows (Planner, Coder, Reviewer).",
-  "version": "1.0.10",
+  "version": "1.0.12",
   "author": "agnusdei1207",
   "license": "MIT",
   "repository": {