neohive 6.4.0 → 6.4.1

package/README.md

@@ -41,6 +41,14 @@
 
 <br />
 
+<p align="center">
+  <img src="https://raw.githubusercontent.com/fakiho/neohive/master/assets/screenshots/hero-animation.gif" alt="Neohive — Terminal + IDE working together" width="100%" />
+</p>
+
+---
+
+<br />
+
 You open Claude Code in one terminal and Gemini CLI in another. Both are powerful — but they can't see each other. You copy context between windows, manually coordinate who does what.
 
 **Neohive removes that bottleneck.** Install once, and your AI agents discover each other, send messages, delegate tasks, review work, and execute multi-step workflows — automatically.
@@ -53,10 +61,12 @@ You open Claude Code in one terminal and Gemini CLI in another. Both are powerfu
 
 - [Quick Start](#-quick-start)
 - [Features](#-features)
+- [Recommended Setup](#-recommended-setup)
 - [How It Works](#-how-it-works)
 - [Supported IDEs & CLIs](#-supported-ides--clis)
 - [Team Templates](#-team-templates)
 - [Dashboard](#-dashboard)
+- [VS Code Extension](#-vs-code-extension)
 - [MCP Tools](#-mcp-tools)
 - [CLI Reference](#%EF%B8%8F-cli-reference)
 - [Configuration](#%EF%B8%8F-configuration)
@@ -77,14 +87,14 @@ That's it. Neohive auto-detects your CLI, writes the MCP config, and creates a `
 
 **MCP config:** `npx neohive init` writes the **absolute path** to the same Node binary that ran the command (so Volta, nvm, or custom installs work even when your IDE’s MCP subprocess has a minimal `PATH`). For **Claude Code**, the project file is `.mcp.json` in the repo root; you can merge the same `neohive` entry into `~/.claude/mcp.json` if you prefer a user-wide setup. Restart the IDE or reload MCP tools after init.
 
-Now open two terminals in the same project:
+Now open two terminals in the same project and paste each prompt into a Claude Code session:
 
 ```
 # Terminal 1
-Register as "Alice" and send a greeting to Bob, then call listen()
+Register as Alice. Send a greeting to Bob, then call listen().
 
 # Terminal 2
-Register as "Bob" and call listen()
+Register as Bob, then call listen().
 ```
 
 Watch them communicate in real time:
@@ -120,6 +130,75 @@ npx neohive dashboard    # opens http://localhost:3000
 
 <br />
 
+## ✅ Recommended Setup
+
+Getting the most out of Neohive takes one extra minute after `init`. Here's what we recommend per tool.
+
+---
+
+### Claude Code
+
+```bash
+npx neohive init --claude
+```
+
+`init` handles MCP config, hooks, and skills in one step. For the smoothest experience:
+
+- **VS Code Extension** — Install the [Neohive extension](https://marketplace.visualstudio.com/items?itemName=alionix.neohive) for automatic MCP setup, in-editor agent status, task board, workflow viewer, and `@neohive` chat participant. The extension configures hooks automatically on activation.
+- **Without the extension** — Run `npx neohive hooks` to install listen-enforcement hooks into `.claude/settings.json`. This keeps agents in the listen loop and prevents them from stopping mid-session. Safe to re-run — your existing hooks are preserved.
+- **Skills** — `init` installs neohive skills and the coordinator agent into `.claude/skills/neohive/`. These teach Claude how to use the MCP tools correctly.
+
+---
+
+<img src="https://raw.githubusercontent.com/fakiho/neohive/master/vscode-extension/assets/screenshots/vscode-extension-overview.png" alt="Neohive VS Code Extension — Team Chat, Agent Sidebar, Task Board, and Workflows" width="100%" />
+
+<br />
+
+### Cursor
+
+```bash
+npx neohive init --cursor
+```
+
+Installs MCP config, skills, commands, and agents into your project's `.cursor/` directory. After init:
+
+- Open Cursor Settings → MCP and **verify that `neohive` is enabled**. Cursor sometimes disables newly added MCP servers by default — toggle it on if needed, then reload.
+- Skills are available as slash commands (e.g. `/neohive-launch-team`, `/neohive-status`).
+
+---
+
+### Antigravity
+
+```bash
+npx neohive init --antigravity
+```
+
+Installs MCP config globally (`~/.gemini/antigravity/mcp_config.json`) and skills into `.agent/skills/neohive/`. After init:
+
+- Open Antigravity Settings → MCP and **verify that `neohive` is enabled**. Like Cursor, Antigravity may disable new MCP servers by default.
+
+---
+
+### Everything at once
+
+```bash
+npx neohive init --all
+```
+
+Configures MCP, hooks, skills, agents, and commands for every detected CLI and IDE in one command.
+
+---
+
+### Troubleshooting
+
+**Agent can't register / MCP tools not found**
+The IDE has likely disabled the neohive MCP server. Restart the IDE first, then go to Settings → MCP (or Tools), find `neohive`, and enable it. After enabling, start a new agent thread — existing sessions won't pick up the newly registered tools.
+
+**Agent stopped listening mid-session**
+Due to a current IDE limitation, agents can occasionally drop out of the listen loop. Simply ask the agent: *"Call listen()"* to resume. We are actively working on a permanent fix.
+
+---
+
 ## 🏗 How It Works
 
 ```
@@ -225,8 +304,8 @@ The MCP server exposes **70+ built-in tools** in one registration list (no separ
 | Category | Tools |
 |----------|-------|
 | **Identity & briefing** | `register` · `list_agents` · `update_profile` · `get_briefing` · `get_guide` |
-| **Messaging** | `send_message` · `broadcast` · `listen` · `listen_group` · `listen_codex` · `wait_for_reply` · `check_messages` · `consume_messages` · `get_notifications` · `ack_message` |
-| **History & search** | `get_history` · `get_summary` · `search_messages` · `get_compressed_history` |
+| **Messaging** | `send_message` · `broadcast` · `listen` · `wait_for_reply` · `messages` |
+| **History & search** | `get_summary` · `get_compressed_history` · `messages` |
 | **Collaboration** | `handoff` · `share_file` · `lock_file` · `unlock_file` |
 | **Tasks** | `create_task` · `update_task` · `list_tasks` |
 | **Workflows** | `create_workflow` · `advance_workflow` · `workflow_status` |
@@ -257,11 +336,12 @@ neohive msg <agent> <text>  # send message from CLI
 neohive doctor              # diagnostic health check
 neohive templates           # list available templates
 neohive hooks               # install listen-enforcement hooks into .claude/settings.json
+neohive skills              # install neohive skills & agents for all detected IDEs
 neohive reset --force       # clear data (auto-archives first)
 neohive uninstall           # remove from all CLI configs
 ```
 
-> **Claude Code users:** Run `npx neohive hooks` after `init` to install listen-enforcement hooks into `.claude/settings.json`. This keeps agents in the listen loop automatically and prevents them from stopping mid-session. Safe to re-run — your existing hooks are preserved.
+> `init` runs `hooks` and `skills` automatically. Run them standalone at any time to update or repair your setup.
 
 <br />
 
@@ -276,6 +356,26 @@ neohive uninstall           # remove from all CLI configs
 
 <br />
 
+## 🧩 VS Code Extension
+
+The [Neohive extension](https://marketplace.visualstudio.com/items?itemName=alionix.neohive) brings agent monitoring and team coordination directly into your editor.
+
+| Feature | Description |
+|---------|-------------|
+| **Agent Sidebar** | See all registered agents, their status (online/stale/offline), and provider in the activity bar |
+| **Task Board** | In-editor kanban board — view and track tasks without opening the dashboard |
+| **Workflow Viewer** | Monitor active workflows and step progress in real time |
+| **`@neohive` Chat** | Query agent status, tasks, and messages directly from Copilot Chat |
+| **Auto MCP Setup** | Configures MCP and hooks automatically on activation — no manual config needed |
+
+**Install:** [VS Code Marketplace](https://marketplace.visualstudio.com/items?itemName=alionix.neohive) — or search "Neohive" in the Extensions panel.
+
+<br />
+
+<img src="https://raw.githubusercontent.com/fakiho/neohive/master/vscode-extension/assets/screenshots/vscode-extension-overview.png" alt="Neohive VS Code Extension — Team Chat, Agent Sidebar, Task Board, and Workflows" width="100%" />
+
+<br />
+
 ## 🔐 Security
 
 Neohive is a **local message broker**. It passes text between CLI terminals via shared files. It does not access the internet, store API keys, or run cloud services.
@@ -341,8 +441,9 @@ node dashboard.js # run the dashboard
 </p>
 
 <p align="center">
+  <a href="https://neohive.alionix.com">Website</a> &middot;
   <a href="https://github.com/fakiho/neohive">GitHub</a> &middot;
   <a href="https://www.npmjs.com/package/neohive">npm</a> &middot;
-  <a href="docs/documentation.md">Docs</a> &middot;
+  <a href="../docs/documentation.md">Docs</a> &middot;
   <a href="mailto:contact@alionix.com">Contact</a>
 </p>

package/cli.js

@@ -47,6 +47,7 @@ function printUsage() {
     npx neohive templates           List available team templates
     npx neohive reset --force       Clear all data (auto-archives first)
     npx neohive hooks               Install listen-enforcement hooks into .claude/settings.json
+    npx neohive skills              Install neohive skills & agents for all detected IDEs
     npx neohive uninstall           Remove from all CLI configs
     npx neohive help                Show this help
 
@@ -134,6 +135,117 @@ function setupClaude(serverPath, cwd) {
 
   fs.writeFileSync(mcpConfigPath, JSON.stringify(mcpConfig, null, 2) + '\n');
   console.log('  [ok] Claude Code: .mcp.json updated');
+  installSkillsForClaude(cwd);
+}
+
+// ─── installSkills ────────────────────────────────────────────────────────────
+// Installs neohive skills, agents, and commands into the appropriate IDE dirs.
+// Safe to re-run: only writes files, never deletes user files outside neohive/.
+
+const SKILLS_SRC = path.join(__dirname, 'neohive-plugin', 'skills');
+const AGENTS_SRC = path.join(__dirname, 'neohive-plugin', 'agents');
+
+function copySkillFile(src, dest, label) {
+  if (!fs.existsSync(src)) return;
+  const existing = fs.existsSync(dest) ? fs.readFileSync(dest, 'utf8') : null;
+  const incoming = fs.readFileSync(src, 'utf8');
+  if (existing === incoming) {
+    console.log(`  [skip] ${label} (unchanged)`);
+  } else {
+    fs.mkdirSync(path.dirname(dest), { recursive: true });
+    fs.writeFileSync(dest, incoming);
+    console.log(`  [ok]   ${label}`);
+  }
+}
+
+function installSkillsForClaude(cwd) {
+  // Project-level skills: .claude/skills/neohive/<skill>/SKILL.md
+  const skillsDir = path.join(cwd, '.claude', 'skills', 'neohive');
+  const agentsDir = path.join(cwd, '.claude', 'skills', 'neohive', 'agents');
+
+  if (!fs.existsSync(SKILLS_SRC)) {
+    console.log('  [skip] Claude skills: source not found (neohive-plugin/skills)');
+    return;
+  }
+
+  for (const skill of fs.readdirSync(SKILLS_SRC)) {
+    const src = path.join(SKILLS_SRC, skill, 'SKILL.md');
+    const dest = path.join(skillsDir, skill, 'SKILL.md');
+    copySkillFile(src, dest, `Claude skill: ${skill}`);
+  }
+
+  if (fs.existsSync(AGENTS_SRC)) {
+    for (const agent of fs.readdirSync(AGENTS_SRC)) {
+      const src = path.join(AGENTS_SRC, agent);
+      const dest = path.join(agentsDir, agent);
+      copySkillFile(src, dest, `Claude agent: ${agent}`);
+    }
+  }
+}
+
+function installSkillsForCursor(cwd) {
+  // .cursor/skills/neohive/<skill>/SKILL.md
+  // .cursor/agents/<agent>.md
+  // .cursor/commands/<command>.md  (derived from skills)
+  if (!fs.existsSync(SKILLS_SRC)) {
+    console.log('  [skip] Cursor skills: source not found (neohive-plugin/skills)');
+    return;
+  }
+
+  for (const skill of fs.readdirSync(SKILLS_SRC)) {
+    const src = path.join(SKILLS_SRC, skill, 'SKILL.md');
+    const dest = path.join(cwd, '.cursor', 'skills', 'neohive', skill, 'SKILL.md');
+    copySkillFile(src, dest, `Cursor skill: ${skill}`);
+
+    // Also install as a command (slash-command style) for user-invocable skills
+    const content = fs.existsSync(src) ? fs.readFileSync(src, 'utf8') : '';
+    if (content.includes('user-invocable: true') || !content.includes('user-invocable: false')) {
+      const cmdDest = path.join(cwd, '.cursor', 'commands', `neohive-${skill}.md`);
+      copySkillFile(src, cmdDest, `Cursor command: neohive-${skill}`);
+    }
+  }
+
+  if (fs.existsSync(AGENTS_SRC)) {
+    for (const agent of fs.readdirSync(AGENTS_SRC)) {
+      const src = path.join(AGENTS_SRC, agent);
+      const dest = path.join(cwd, '.cursor', 'agents', `neohive-${agent}`);
+      copySkillFile(src, dest, `Cursor agent: neohive-${agent}`);
+    }
+  }
+}
+
+function installSkillsForAntigravity(cwd) {
+  // .agent/skills/neohive/<skill>/SKILL.md  (already partially done in setupAntigravity)
+  if (!fs.existsSync(SKILLS_SRC)) {
+    console.log('  [skip] Antigravity skills: source not found (neohive-plugin/skills)');
+    return;
+  }
+
+  for (const skill of fs.readdirSync(SKILLS_SRC)) {
+    const src = path.join(SKILLS_SRC, skill, 'SKILL.md');
+    const dest = path.join(cwd, '.agent', 'skills', 'neohive', skill, 'SKILL.md');
+    copySkillFile(src, dest, `Antigravity skill: ${skill}`);
+  }
+
+  if (fs.existsSync(AGENTS_SRC)) {
+    for (const agent of fs.readdirSync(AGENTS_SRC)) {
+      const src = path.join(AGENTS_SRC, agent);
+      const dest = path.join(cwd, '.agent', 'agents', `neohive-${agent}`);
+      copySkillFile(src, dest, `Antigravity agent: neohive-${agent}`);
+    }
+  }
+}
+
+function installSkills(targets, cwd) {
+  console.log('');
+  console.log('  Installing Neohive skills & agents...');
+  for (const target of targets) {
+    switch (target) {
+      case 'claude':      installSkillsForClaude(cwd);      break;
+      case 'cursor':      installSkillsForCursor(cwd);      break;
+      case 'antigravity': installSkillsForAntigravity(cwd); break;
+    }
+  }
 }
 
 function setupGemini(serverPath, cwd) {
@@ -447,15 +559,7 @@ function setupAntigravity(cwd) {
 
   fs.writeFileSync(mcpPath, JSON.stringify(config, null, 2) + '\n');
   console.log('  [ok] Antigravity: ~/.gemini/antigravity/mcp_config.json updated');
-
-  // Write skill
-  const skillDir = path.join(cwd, '.agent', 'skills', 'neohive');
-  if (!fs.existsSync(skillDir)) fs.mkdirSync(skillDir, { recursive: true });
-  const skillPath = path.join(skillDir, 'SKILL.md');
-  if (!fs.existsSync(skillPath)) {
-    fs.writeFileSync(skillPath, neohiveAgentRules('Gemini'));
-    console.log('  [ok] Antigravity: .agent/skills/neohive/SKILL.md created');
-  }
+  installSkillsForAntigravity(cwd);
 }
 
 function neohiveAgentRules(defaultName) {
@@ -563,6 +667,7 @@ function setupCursor(serverPath, cwd) {
 
   fs.writeFileSync(mcpConfigPath, JSON.stringify(mcpConfig, null, 2) + '\n');
   console.log('  [ok] Cursor IDE: .cursor/mcp.json updated');
+  installSkillsForCursor(cwd);
 }
 
 // Setup Ollama agent bridge script
@@ -765,6 +870,9 @@ function init() {
     }
   }
 
+  // Auto-install hooks for Claude targets
+  if (targets.includes('claude')) installHooks();
+
   // Add .neohive/ and MCP config files to .gitignore
   const gitignoreEntries = ['.neohive/', '.mcp.json', '.cursor/mcp.json', '.codex/', '.gemini/'];
   if (fs.existsSync(gitignorePath)) {
@@ -818,12 +926,6 @@ function init() {
     console.log('    npx neohive status');
     console.log('    npx neohive doctor');
     console.log('');
-    if (targets.includes('claude')) {
-      console.log('  \x1b[33m  Tip (Claude Code):\x1b[0m Run `npx neohive hooks` to install listen-enforcement');
-      console.log('  hooks into .claude/settings.json. Keeps agents in the listen loop automatically.');
-      console.log('  Your existing hooks will not be removed.');
-      console.log('');
-    }
   }
 }
 
@@ -1504,6 +1606,9 @@ switch (command) {
   case 'hooks':
     installHooks();
     break;
+  case 'skills':
+    installSkills(detectCLIs().length ? detectCLIs() : ['claude', 'cursor', 'antigravity'], process.cwd());
+    break;
   case 'uninstall':
   case 'remove':
     uninstall();

package/dashboard.html

@@ -4,18 +4,6 @@
 <meta charset="UTF-8">
 <meta name="viewport" content="width=device-width, initial-scale=1.0">
 <title>Neohive | Multi-Agent Coordination Dashboard</title>
-<meta name="description" content="Neohive — The ultimate multi-agent coordination and management dashboard. Orchestrate, monitor, and scale your AI workforce in real-time.">
-<meta name="keywords" content="AI, multi-agent systems, orchestration, dashboard, neohive, agents, automation">
-<meta name="author" content="Neohive Team">
-<meta property="og:title" content="Neohive | Multi-Agent Coordination Dashboard">
-<meta property="og:description" content="Orchestrate and scale your autonomous agent workforce with a stunning, high-performance real-time dashboard.">
-<meta property="og:type" content="website">
-<meta property="og:image" content="https://neohive.ai/og-image.png">
-<meta name="twitter:card" content="summary_large_image">
-<meta name="twitter:title" content="Neohive | Multi-Agent Coordination Dashboard">
-<meta name="twitter:description" content="The command center for your AI agent workforce. Real-time monitoring and scalable orchestration.">
-<meta name="twitter:image" content="https://neohive.ai/og-image.png">
-<meta name="robots" content="index, follow">
 <link rel="icon" href="favicon.png" type="image/png" sizes="16x16">
 <link rel="icon" href="logo.svg" type="image/svg+xml">
 <link rel="preconnect" href="https://fonts.googleapis.com">
@@ -4603,7 +4591,7 @@
   </div>
 </div>
 <div class="app-footer">
-  <span>Neohive v6.0.0</span>
+  <span>Neohive v6.4.1</span>
 </div>
 <div class="profile-popup" id="profile-popup" onclick="event.stopPropagation()">
   <div class="profile-popup-header">
@@ -5294,10 +5282,10 @@ function respawnAgent(agentName) {
 }
 
 function generateFallbackRespawnPrompt(agentName) {
-  return 'Register as \'' + agentName + '\', then call get_briefing() and listen_group() to rejoin the conversation. ' +
+  return 'Register as \'' + agentName + '\', then call get_briefing() and listen() to rejoin the conversation. ' +
     'You are resuming a previous session — call get_compressed_history() to catch up on what you missed. ' +
     'Check your workspace with workspace_read() for any saved state. ' +
-    'Then call listen_group() and respond to any pending messages.';
+    'Then call listen() and respond to any pending messages.';
 }
 
 function showRespawnModal(agentName, prompt) {
@@ -5503,7 +5491,7 @@ function renderMessages(messages) {
     } else {
       el.innerHTML = '<div class="empty-state">' +
         '<div class="empty-icon" style="font-size:40px;opacity:0.2">--</div>' +
-        '<div class="empty-text">Neohive v6.0</div>' +
+        '<div class="empty-text">Neohive v6.4</div>' +
         '<div class="empty-sub">Autonomous AI agent teams — one command, zero babysitting</div>' +
         '<div class="onboard-steps">' +
           '<div class="onboard-step" style="margin-bottom:12px"><span class="onboard-num" style="background:var(--accent)"></span><span style="font-weight:600">Quickest start — one command:</span></div>' +
@@ -10313,7 +10301,7 @@ function doLaunch() {
   }
 
   // Use generated prompt if available, otherwise build a default
-  var launchPrompt = window._generatedLaunchPrompt || 'You are agent "' + agentName + '". Use the register tool to register as "' + agentName + '", then use listen_group() to join the conversation.';
+  var launchPrompt = window._generatedLaunchPrompt || 'You are agent "' + agentName + '". Use the register tool to register as "' + agentName + '", then use listen() to join the conversation.';
   navigator.clipboard.writeText(launchPrompt).catch(function() {});
   selectedCli = cli;
 
@@ -10404,7 +10392,7 @@ function doLaunchAll() {
     w.state = 'launching';
     renderLaunchStatusTracker();
 
-    var agentPrompt = w.prompt || 'You are agent "' + w.name + '". Use the register tool to register as "' + w.name + '", then use listen_group() to join the conversation.';
+    var agentPrompt = w.prompt || 'You are agent "' + w.name + '". Use the register tool to register as "' + w.name + '", then use listen() to join the conversation.';
 
     lttFetch('/api/launch', {
       method: 'POST',
@@ -10533,7 +10521,7 @@ function renderDocs() {
   el.innerHTML =
     '<div class="docs-container">' +
 
-    '<h2>Neohive v6.0</h2>' +
+    '<h2>Neohive v6.4</h2>' +
     '<p class="docs-subtitle">True Autonomy Engine \u2014 AI agents that self-organize, self-verify, and never stop working. Works with Claude Code, Gemini CLI, Codex CLI, and Cursor IDE.</p>' +
 
     // Quick Start — One Command
@@ -10571,7 +10559,7 @@ function renderDocs() {
       '<p>Opens this web dashboard at <code>http://localhost:3777</code>. You can watch agents chat in real-time, send them messages, and manage your team.</p>' +
       '<h4>3. Start Your Agents</h4>' +
       '<p>Open two or more terminal windows in your project folder. In each one, start your AI CLI (e.g. type <code>claude</code>) and tell it to register:</p>' +
-      '<pre><code># Terminal 1\nRegister as "Alice" and use listen_group() to join the conversation.\n\n# Terminal 2\nRegister as "Bob" and use listen_group() to join the conversation.</code></pre>' +
+      '<pre><code># Terminal 1\nRegister as "Alice" and use listen() to join the conversation.\n\n# Terminal 2\nRegister as "Bob" and use listen() to join the conversation.</code></pre>' +
       '<p>That\'s it! Your agents can now talk to each other. Or use the <strong>Launch</strong> tab to do this with one click.</p>' +
     '</div>' +
 
@@ -10655,7 +10643,7 @@ function renderDocs() {
       '<h4>Can I run multiple projects?</h4>' +
       '<p>Yes. Each project has its own <code>.neohive/</code> directory. The dashboard supports multiple projects \u2014 click the project selector in the header to switch between them.</p>' +
       '<h4>How do I send a message to agents from the dashboard?</h4>' +
-      '<p>Click any agent\'s avatar in the Messages tab. A dialog lets you type and send a message that agents will receive on their next <code>listen_group()</code> call.</p>' +
+      '<p>Click any agent\'s avatar in the Messages tab. A dialog lets you type and send a message that agents will receive on their next <code>listen()</code> call.</p>' +
     '</div>' +
 
     '</div>';
@@ -10674,8 +10662,8 @@ function copyLaunchPrompt() {
   if (!prompt) {
     var agentName = document.getElementById('launch-name').value.trim();
     prompt = agentName
-      ? 'You are agent "' + agentName + '". Use the register tool to register as "' + agentName + '", then use listen_group() to join the conversation.'
-      : 'Register with the neohive MCP tools and use listen_group() to join the conversation.';
+      ? 'You are agent "' + agentName + '". Use the register tool to register as "' + agentName + '", then use listen() to join the conversation.'
+      : 'Register with the neohive MCP tools and use listen() to join the conversation.';
   }
   navigator.clipboard.writeText(prompt).then(function() {
     var resultEl = document.getElementById('launch-result');

package/dashboard.js

@@ -2662,9 +2662,9 @@ const server = http.createServer(async (req, res) => {
       prompt += `**Instructions:**\n`;
       prompt += `1. Register as "${agentName}" using the register tool\n`;
       prompt += `2. Call get_briefing() for full project context\n`;
-      prompt += `3. Call listen_group() to rejoin the conversation\n`;
+      prompt += `3. Call listen() to rejoin the conversation\n`;
       prompt += `4. Announce you're back and pick up your active tasks\n`;
-      prompt += `5. Stay in listen_group() loop — never stop listening\n`;
+      prompt += `5. Stay in listen() loop — never stop listening\n`;
 
       res.writeHead(200, { 'Content-Type': 'application/json' });
       res.end(JSON.stringify({
@@ -2717,7 +2717,7 @@ const server = http.createServer(async (req, res) => {
         try {
           const messagesFile = filePath('messages.jsonl', projectPath);
           const historyFile = filePath('history.jsonl', projectPath);
-          const modeText = newMode === 'responsive' ? 'Coordinator stays with human, uses consume_messages().' : 'Coordinator runs autonomously in listen() loop.';
+          const modeText = newMode === 'responsive' ? 'Coordinator stays with human, uses messages(action="check").' : 'Coordinator runs autonomously in listen() loop.';
           const sysMsg = { id: Date.now().toString(36) + Math.random().toString(36).slice(2, 8), from: '__system__', to: '__group__', content: `[MODE] Coordinator mode changed to "${newMode}". ${modeText} Coordinator: call get_guide() to update your instructions.`, timestamp: new Date().toISOString(), system: true };
           fs.appendFileSync(messagesFile, JSON.stringify(sysMsg) + '\n');
           fs.appendFileSync(historyFile, JSON.stringify(sysMsg) + '\n');

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "neohive",
-  "version": "6.4.0",
+  "version": "6.4.1",
   "description": "The MCP collaboration layer for AI CLI tools. Turn Claude Code, Gemini CLI, and Codex CLI into a team.",
   "main": "server.js",
   "bin": {

package/server.js

@@ -1183,7 +1183,7 @@ function buildGuide(level = 'standard') {
     } else {
       rules.push('ROLE: Managed agent. The manager controls your turn.');
       rules.push('LOOP: listen() → receive work → update_task(id, "in_progress") → do work → update_task(id, "done") → send_message(manager, summary) → listen(). Never stop.');
-      rules.push('Never call get_work() or check_messages() in managed mode.');
+      rules.push('Never call get_work() or messages() in managed mode.');
     }
     rules.push('Keep messages short (2-3 paragraphs). Report what you did and what files changed.');
   }
@@ -1260,16 +1260,16 @@ function buildGuide(level = 'standard') {
   if (isLeadRole) {
     const coordinatorMode = getConfig().coordinator_mode || 'responsive';
     if (coordinatorMode === 'responsive') {
-      rules.push('COORDINATOR: Use consume_messages() to check updates non-blockingly. Do NOT block in listen() — stay responsive to the user.');
+      rules.push('COORDINATOR: Use messages(action="check") to check updates non-blockingly. Do NOT block in listen() — stay responsive to the user.');
     } else {
       rules.push('COORDINATOR: Use listen() to wait for agent results. Only return to human when all tasks are done or blocked.');
     }
     rules.push('Coordinators do NOT edit files or write code. Delegate ALL code work to other agents.');
   }
 
-  const listenCmd = isManagedMode() ? 'listen()' : (mode === 'group' ? 'listen_group()' : 'listen()');
+  const listenCmd = isManagedMode() ? 'listen()' : (mode === 'group' ? 'listen(mode="group")' : 'listen()');
   if (!isLeadRole) {
-    rules.push(`After EVERY action, call ${listenCmd}. Never use sleep() or poll with check_messages().`);
+    rules.push(`After EVERY action, call ${listenCmd}. Never use sleep() or poll with messages().`);
   }
 
   if (level === 'minimal') {
@@ -1570,7 +1570,7 @@ function toolRegister(name, provider = null, skills = null) {
       const coordinatorMode = getConfig().coordinator_mode || 'responsive';
       nextAction = coordinatorMode === 'autonomous'
         ? 'Call get_briefing() to load project context, then listen() to coordinate your team.'
-        : 'Call get_briefing() to load project context, then consume_messages() to check for pending work.';
+        : 'Call get_briefing() to load project context, then messages(action="check") to check for pending work.';
     }
 
     // --- Build the result: next_action FIRST, then context ---
@@ -1695,7 +1695,7 @@ async function toolSendMessage(content, to = null, reply_to = null, channel = nu
   const effectiveSendLimit = isAutonomousMode() ? 5 : sendLimit;
   const myRole = (getProfiles()[registeredName] || {}).role;
   if (isGroupMode() && sendsSinceLastListen >= effectiveSendLimit && myRole !== 'Coordinator') {
-    return { error: `You must call listen_group() before sending again. You've sent ${sendsSinceLastListen} message(s) without listening (limit: ${effectiveSendLimit}). This prevents message storms.` };
+    return { error: `You must call listen() before sending again. You've sent ${sendsSinceLastListen} message(s) without listening (limit: ${effectiveSendLimit}). This prevents message storms.` };
   }
 
   // Response budget: track unaddressed sends, hint when depleted
@@ -1993,7 +1993,7 @@ async function toolSendMessage(content, to = null, reply_to = null, channel = nu
   if (!recipientAlive) {
     result.warning = `Agent "${to}" appears offline (PID not running). Message queued but may not be received until they reconnect.`;
   } else if (to !== '__user__' && agents[to] && !agents[to].listening_since) {
-    result.note = `Agent "${to}" is currently working (not in listen mode). Message queued — they'll see it when they finish their current task and call listen_group().`;
+    result.note = `Agent "${to}" is currently working (not in listen mode). Message queued — they'll see it when they finish their current task and call listen().`;
   }
 
   // Coordinator enforcement: warn if sending work assignment without creating a task first
@@ -2036,7 +2036,7 @@ function toolBroadcast(content) {
   const effectiveSendLimitBcast = isAutonomousMode() ? 5 : sendLimit;
   const myRole = (getProfiles()[registeredName] || {}).role;
   if (isGroupMode() && sendsSinceLastListen >= effectiveSendLimitBcast && myRole !== 'Coordinator') {
-    return { error: `You must call listen_group() before broadcasting again. You've sent ${sendsSinceLastListen} message(s) without listening (limit: ${effectiveSendLimitBcast}).` };
+    return { error: `You must call listen() before broadcasting again. You've sent ${sendsSinceLastListen} message(s) without listening (limit: ${effectiveSendLimitBcast}).` };
   }
 
   const rateErr = checkRateLimit(content, '__broadcast__');
@@ -2355,10 +2355,10 @@ async function toolListenCodex(from = null, outcome = null, task_id = null, summ
     const taskList = getTasks();
     const task = taskList.find(t => t.id === task_id);
     if (!task) {
-      return { error: true, message: `Invalid task_id "${task_id}" — task does not exist. Check list_tasks() and call listen_codex() again with the correct task_id.` };
+      return { error: true, message: `Invalid task_id "${task_id}" — task does not exist. Check list_tasks() and call listen(mode="codex") again with the correct task_id.` };
     }
     if (task.assignee && task.assignee !== registeredName) {
-      return { error: true, message: `Task "${task_id}" is assigned to ${task.assignee}, not to you (${registeredName}). You cannot update another agent's task via listen_codex().` };
+      return { error: true, message: `Task "${task_id}" is assigned to ${task.assignee}, not to you (${registeredName}). You cannot update another agent's task via listen(mode="codex").` };
     }
     const statusMap = { completed: 'done', blocked: 'blocked', failed: 'blocked_permanent' };
     const newStatus = statusMap[outcome];
@@ -2486,9 +2486,9 @@ function toolSetConversationMode(mode) {
   }
 
   const messages = {
-    group: 'Group mode enabled. Use listen_group() to receive batched messages. All messages are shared with everyone.',
+    group: 'Group mode enabled. Use listen(mode="group") to receive batched messages. All messages are shared with everyone.',
     direct: 'Direct mode enabled. Use listen() for point-to-point messaging.',
-    managed: 'Managed mode enabled. Call claim_manager() to become the manager, or wait for the manager to give you the floor via yield_floor(). Use listen() or listen_group() to receive messages.',
+    managed: 'Managed mode enabled. Call claim_manager() to become the manager, or wait for the manager to give you the floor via yield_floor(). Use listen() to receive messages.',
   };
   return { success: true, mode, message: messages[mode] };
 }
@@ -2649,10 +2649,10 @@ async function toolListenGroup(outcome = null, task_id = null, summary = null) {
     const taskList = getTasks();
     const task = taskList.find(t => t.id === task_id);
     if (!task) {
-      return { error: true, message: `Invalid task_id "${task_id}" — task does not exist. Check list_tasks() and call listen_group() again with the correct task_id.` };
+      return { error: true, message: `Invalid task_id "${task_id}" — task does not exist. Check list_tasks() and call listen() again with the correct task_id.` };
     }
     if (task.assignee && task.assignee !== registeredName) {
-      return { error: true, message: `Task "${task_id}" is assigned to ${task.assignee}, not to you (${registeredName}). You cannot update another agent's task via listen_group().` };
+      return { error: true, message: `Task "${task_id}" is assigned to ${task.assignee}, not to you (${registeredName}). You cannot update another agent's task via listen().` };
     }
     const statusMap = { completed: 'done', blocked: 'blocked', failed: 'blocked_permanent' };
     const newStatus = statusMap[outcome];
@@ -2906,8 +2906,8 @@ function classifyPriority(msg) {
   return 'normal';
 }
 
-// Build the response for listen_group — kept lean to reduce context accumulation
-// Context/history removed: agents should call get_history() when they need it
+// Build the response for listen (group mode) — kept lean to reduce context accumulation
+// Context/history removed: agents should call messages(action="history") when they need it
 function buildListenGroupResponse(batch, consumed, agentName, listenStart) {
   saveConsumedIds(agentName, consumed);
   touchActivity();
@@ -5465,7 +5465,7 @@ function toolStartPlan(params) {
   broadcastSystemMessage(
     `[PLAN LAUNCHED] "${name}" — ${steps.length} steps, autonomous mode, ${useParallel ? 'parallel' : 'sequential'}. ` +
     `${startedSteps.length} step(s) started. ` +
-    `All agents: call get_work() to enter the autonomous work loop. Do NOT call listen_group().`
+    `All agents: call get_work() to enter the autonomous work loop. Do NOT call listen().`
   );
 
   touchActivity();
@@ -5886,7 +5886,7 @@ function triggerStandupIfDue() {
     if (inProgress.length > 0) summary += ` In progress: ${inProgress.map(t => `"${t.title}" (${t.assignee || '?'})`).join(', ')}.`;
     if (blocked.length > 0) summary += ` BLOCKED: ${blocked.map(t => `"${t.title}" (${t.assignee || '?'})`).join(', ')}.`;
     if (recentDone.length > 0) summary += ` Recently done: ${recentDone.length} task(s).`;
-    summary += ' Each agent: report what you did, what\'s blocked, what\'s next. Then call listen_group().';
+    summary += ' Each agent: report what you did, what\'s blocked, what\'s next. Then call listen().';
 
     broadcastSystemMessage(summary, registeredName);
   } catch (e) { log.warn("standup trigger failed:", e.message); }
@@ -6180,7 +6180,7 @@ function toolGetGuide(level = 'standard') {
   const guide = buildGuide(level);
   guide.your_name = registeredName;
   if (level !== 'minimal') {
-    guide.workflow = '1. get_briefing → 2. list_tasks/suggest_task → 3. claim task → 4. lock_file → 5. work → 6. unlock_file → 7. update_task done → 8. listen_group';
+    guide.workflow = '1. get_briefing → 2. list_tasks/suggest_task → 3. claim task → 4. lock_file → 5. work → 6. unlock_file → 7. update_task done → 8. listen()';
   }
   return guide;
 }
@@ -7191,7 +7191,7 @@ server.setRequestHandler(ListToolsRequestSchema, async () => {
     tools: [
       {
         name: 'register',
-        description: 'Register this agent\'s identity. Must be called first. Returns a collaboration guide with all tool categories, critical rules, and workflow patterns — READ IT CAREFULLY before doing anything else. Then call get_briefing() for project context, then listen_group() to join the conversation.',
+        description: 'Register this agent\'s identity. Must be called first. Returns a collaboration guide with all tool categories, critical rules, and workflow patterns — READ IT CAREFULLY before doing anything else. Then call get_briefing() for project context, then listen() to join the conversation.',
         inputSchema: {
           type: 'object',
           properties: {
@@ -7289,7 +7289,7 @@ server.setRequestHandler(ListToolsRequestSchema, async () => {
       },
       {
         name: 'listen',
-        description: 'Listen for messages. Use mode="standard" (default, direct 1:1), mode="group" (group/managed conversation, batched), or mode="codex" (Codex CLI — returns after 90s). Auto-detects mode from conversation state when mode is omitted. Replaces listen_group and listen_codex (now deprecated aliases).',
+        description: 'Listen for messages. Use mode="standard" (default, direct 1:1), mode="group" (group/managed conversation, batched), or mode="codex" (Codex CLI — returns after 90s). Auto-detects mode from conversation state when mode is omitted.',
         inputSchema: {
           type: 'object',
           properties: {
@@ -7901,7 +7901,7 @@ server.setRequestHandler(CallToolRequestSchema, async (request) => {
         try {
           const pending = getUnconsumedMessages(registeredName);
           const pendingHint = pending.length > 0
-            ? `${pending.length} agent update(s) waiting. Call consume_messages() to read them.`
+            ? `${pending.length} agent update(s) waiting. Call messages(action="consume") to read them.`
             : null;
           if (!na || bareListenRe.test(na)) {
             // No guidance or bare listen() — replace with coordinator hint or nothing