docs: update landing page for v0.3.5 release

Bump Docker image tags to 0.3.5 (API and frontend). Update MCP docs
(EN/ES) to reflect runtime status monitoring: statuses changed from
Configured/Error to Running/Failed to match the new real-time status
reporting parsed from Claude Code init events and OpenCode runtime events.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Author: barckcode

docker-compose.yml

@@ -5,7 +5,7 @@ services:
     command: ["--auth", "${NATS_AUTH_TOKEN}", "-m", "8222"]
 
   api:
-    image: ghcr.io/helmcode/agent_crew_api:0.3.3
+    image: ghcr.io/helmcode/agent_crew_api:0.3.5
     restart: unless-stopped
     user: root
     ports:
@@ -29,7 +29,7 @@ services:
       - agentcrew_db:/data
 
   frontend:
-    image: ghcr.io/helmcode/agent_crew:0.3.2
+    image: ghcr.io/helmcode/agent_crew:0.3.5
     restart: unless-stopped
     ports:
       - "${FRONTEND_PORT:-8080}:8080"

web/src/pages/docs/mcp.astro

@@ -104,7 +104,11 @@ import DocsLayout from '../../layouts/DocsLayout.astro';
     <h2>MCP Status Monitoring</h2>
 
     <p>
-      Each MCP server shows its status in the Settings modal:
+      Each MCP server shows its runtime status in the Settings modal.
+      AgentCrew reports the actual runtime state of each MCP server by
+      parsing status events from the AI provider (Claude Code init events
+      and OpenCode runtime events). This gives you real-time visibility
+      into whether each server started successfully or failed.
     </p>
 
     <table>
@@ -117,32 +121,34 @@ import DocsLayout from '../../layouts/DocsLayout.astro';
       </thead>
       <tbody>
         <tr>
-          <td><strong>Configured</strong></td>
+          <td><strong>Running</strong></td>
           <td>Green dot</td>
-          <td>The server configuration was written successfully and passed health checks.</td>
+          <td>The MCP server is running and available to agents. Confirmed by the AI provider runtime.</td>
         </tr>
         <tr>
-          <td><strong>Error</strong></td>
+          <td><strong>Failed</strong></td>
           <td>Red dot</td>
-          <td>The server failed health checks. Click the arrow to expand error details.</td>
+          <td>The MCP server failed to start. Click the arrow to expand error details reported by the runtime.</td>
         </tr>
         <tr>
           <td><strong>Pending</strong></td>
           <td>Gray dot</td>
-          <td>The server is waiting for status from the agent container.</td>
+          <td>The server is waiting for runtime status from the agent container. This is the initial state before the AI provider reports back.</td>
         </tr>
       </tbody>
     </table>
 
     <p>
-      When an MCP server has an error, the error details are expandable
-      directly in the server list. Common causes include:
+      When an MCP server has a <strong>Failed</strong> status, the error
+      details are expandable directly in the server list. Common causes
+      include:
     </p>
 
     <ul>
       <li>Package not found (wrong npm/pip package name)</li>
       <li>Remote server unreachable (wrong URL or network issue)</li>
       <li>Command not found (binary not installed in the agent image)</li>
+      <li>Authentication failure (invalid credentials or expired tokens)</li>
       <li>Timeout during dependency download</li>
     </ul>
 

web/src/pages/es/docs/mcp.astro

@@ -105,7 +105,11 @@ import DocsLayout from '../../../layouts/DocsLayout.astro';
     <h2>Monitoreo de estado MCP</h2>
 
     <p>
-      Cada servidor MCP muestra su estado en el modal de Settings:
+      Cada servidor MCP muestra su estado de runtime en el modal de Settings.
+      AgentCrew reporta el estado real de ejecucion de cada servidor MCP
+      parseando eventos de estado del proveedor de IA (eventos init de
+      Claude Code y eventos de runtime de OpenCode). Esto te da visibilidad
+      en tiempo real sobre si cada servidor inicio correctamente o fallo.
     </p>
 
     <table>
@@ -118,32 +122,34 @@ import DocsLayout from '../../../layouts/DocsLayout.astro';
       </thead>
       <tbody>
         <tr>
-          <td><strong>Configured</strong></td>
+          <td><strong>Running</strong></td>
           <td>Punto verde</td>
-          <td>La configuracion del servidor se escribio correctamente y paso los health checks.</td>
+          <td>El servidor MCP esta ejecutandose y disponible para los agentes. Confirmado por el runtime del proveedor de IA.</td>
         </tr>
         <tr>
-          <td><strong>Error</strong></td>
+          <td><strong>Failed</strong></td>
           <td>Punto rojo</td>
-          <td>El servidor fallo los health checks. Haz clic en la flecha para expandir los detalles del error.</td>
+          <td>El servidor MCP fallo al iniciar. Haz clic en la flecha para expandir los detalles del error reportados por el runtime.</td>
         </tr>
         <tr>
           <td><strong>Pending</strong></td>
           <td>Punto gris</td>
-          <td>El servidor esta esperando el estado del contenedor del agente.</td>
+          <td>El servidor esta esperando el estado de runtime del contenedor del agente. Este es el estado inicial antes de que el proveedor de IA reporte.</td>
         </tr>
       </tbody>
     </table>
 
     <p>
-      Cuando un servidor MCP tiene un error, los detalles del error son expandibles
-      directamente en la lista de servidores. Las causas comunes incluyen:
+      Cuando un servidor MCP tiene estado <strong>Failed</strong>, los
+      detalles del error son expandibles directamente en la lista de
+      servidores. Las causas comunes incluyen:
     </p>
 
     <ul>
       <li>Paquete no encontrado (nombre incorrecto del paquete npm/pip)</li>
       <li>Servidor remoto inalcanzable (URL incorrecta o problema de red)</li>
       <li>Comando no encontrado (binario no instalado en la imagen del agente)</li>
+      <li>Fallo de autenticacion (credenciales invalidas o tokens expirados)</li>
       <li>Timeout durante la descarga de dependencias</li>
     </ul>