---
title: Construye un modo de orquestación
url: https://platform.claude.com/docs/es/build-with-claude/mid-conversation-effort-example
description: Construye un modo a nivel de sesión que otorga consentimiento permanente para el fan-out multiagente, activado y desactivado con mensajes del sistema a mitad de conversación.
---

Un "orchestration mode" (modo de orquestación) es un interruptor a nivel de sesión. Cuando está activado, el modelo aplica la máxima minuciosidad a cada solicitud sustancial: primero explora la tarea por sí mismo y luego distribuye el trabajo entre "subagents" (subagentes) en paralelo de forma predeterminada. Cuando está desactivado, la misma herramienta de orquestación vuelve a requerir una activación explícita en cada solicitud.

El modo no es un parámetro de la API. Se construye por completo a partir de piezas documentadas:

1. **Un nivel de esfuerzo:** las solicitudes se ejecutan con un valor documentado de [esfuerzo](https://platform.claude.com/docs/es/build-with-claude/effort), como `xhigh`. No existe ningún nivel oculto por encima de los que aparecen en esa página. Este ejemplo establece el esfuerzo en el nivel superior de cada solicitud, lo cual no requiere ningún encabezado beta.
2. **Un recordatorio de modo:** un [mensaje del sistema a mitad de conversación](https://platform.claude.com/docs/es/build-with-claude/mid-conversation-system-messages) le indica al modelo que el modo está activo. Además, se envía un recordatorio de una línea cada cierto número de turnos y un aviso de salida cuando el modo se desactiva. El campo `system` de nivel superior nunca cambia, por lo que el prefijo almacenado en caché permanece intacto.
3. **Consentimiento permanente en la descripción de la herramienta:** la descripción de la herramienta de orquestación establece que, mientras el modo esté activado, el modelo debe redactar y ejecutar un flujo de trabajo para cada tarea sustancial sin preguntar primero.

<Note>
  Este ejemplo usa mensajes del sistema a mitad de conversación. Para conocer los modelos y plataformas que los admiten, consulta [Mensajes del sistema a mitad de conversación](https://platform.claude.com/docs/es/build-with-claude/mid-conversation-system-messages). El "fan-out" (distribución en paralelo) multiplica el uso de tokens, ya que una sola solicitud puede generar muchas conversaciones de subagentes. Reserva el modo para trabajos que justifiquen el costo.
</Note>

## Configura el bucle

El ejemplo es un solo archivo. Las constantes controlan el nivel de esfuerzo, la forma de la distribución en paralelo y la frecuencia con la que se vuelve a enviar el recordatorio del modo. `MAX_CONCURRENT` limita cuántos subagentes se ejecutan al mismo tiempo (la versión en PHP es secuencial y la ignora). `MAX_TOTAL_SUBTASKS` limita cuántos subagentes puede poner en cola el modelo en una sola llamada a Workflow. Separar ambos límites permite que el modelo planifique una gran cantidad de trabajo pendiente sin lanzarlo todo a la vez.

La comprobación de `DOC_TEST_MODE` limita los bucles a un solo turno cuando esa variable de entorno está definida. Así, el sistema automatizado de pruebas de la documentación puede validar que el archivo compila y termina rápidamente sin ejecutar la orquestación completa. Déjala sin definir cuando ejecutes el ejemplo tú mismo.

<CodeGroup exclude="shell:cURL, shell:CLI">
  ```python Python
  import atexit
  import concurrent.futures
  import hashlib
  import json
  import os
  import shutil
  import subprocess
  import sys
  import tempfile
  import threading

  import anthropic

  client = anthropic.Anthropic()

  MODEL = "claude-opus-5-5"
  EFFORT = "xhigh"

  SYSTEM_PROMPT = (
      "You are a helpful general-purpose agent. Answer the user's request directly."
  )

  REQUEST_TIMEOUT_SECONDS = 600
  BASH_TIMEOUT_SECONDS = 60
  TOOL_RESULT_MAX_CHARS = 8000
  MAX_CONCURRENT = 10
  DOC_TEST_MODE = bool(os.environ.get("DOC_TEST_MODE"))
  MAX_TOTAL_SUBTASKS = 2 if DOC_TEST_MODE else 200
  MAX_SUBAGENT_TURNS = 1 if DOC_TEST_MODE else 15
  MAX_MAIN_TURNS = 1 if DOC_TEST_MODE else 30
  TURNS_BETWEEN_REFRESHERS = 10
  JOURNAL_PATH = os.environ.get("ORCH_JOURNAL") or "orchestration_journal.json"
  ```

  ```typescript TypeScript
  import { exec } from "node:child_process";
  import { createHash } from "node:crypto";
  import { mkdtempDisposable, readFile, rename, writeFile } from "node:fs/promises";
  import { tmpdir } from "node:os";
  import { join } from "node:path";

  import Anthropic from "@anthropic-ai/sdk";

  const client = new Anthropic();

  const MODEL = "claude-opus-5-5";
  const EFFORT = "xhigh";

  const SYSTEM_PROMPT =
    "You are a helpful general-purpose agent. Answer the user's request directly.";

  const REQUEST_TIMEOUT_SECONDS = 600;
  const BASH_TIMEOUT_SECONDS = 60;
  const TOOL_RESULT_MAX_CHARS = 8000;
  const MAX_CONCURRENT = 10;
  const DOC_TEST_MODE = Boolean(process.env.DOC_TEST_MODE);
  const MAX_TOTAL_SUBTASKS = DOC_TEST_MODE ? 2 : 200;
  const MAX_SUBAGENT_TURNS = DOC_TEST_MODE ? 1 : 15;
  const MAX_MAIN_TURNS = DOC_TEST_MODE ? 1 : 30;
  const TURNS_BETWEEN_REFRESHERS = 10;
  const JOURNAL_PATH = process.env.ORCH_JOURNAL || "orchestration_journal.json";
  ```

  ```csharp C#
  using System.Diagnostics;
  using System.Security.Cryptography;
  using System.Text;
  using System.Text.Json;
  using Anthropic;
  using Anthropic.Models.Messages;

  AnthropicClient client = new();

  const Model model = Model.ClaudeOpus5_5;
  var effort = Effort.Xhigh;

  const string systemPrompt = "You are a helpful general-purpose agent. Answer the user's request directly.";

  const int requestTimeoutSeconds = 600;
  const int bashTimeoutSeconds = 60;
  const int toolResultMaxChars = 8000;
  const int maxConcurrent = 10;
  var docTestMode = Environment.GetEnvironmentVariable("DOC_TEST_MODE") is { Length: > 0 };
  var maxTotalSubtasks = docTestMode ? 2 : 200;
  var maxSubagentTurns = docTestMode ? 1 : 15;
  var maxMainTurns = docTestMode ? 1 : 30;
  const int turnsBetweenRefreshers = 10;
  var journalPath = Environment.GetEnvironmentVariable("ORCH_JOURNAL") is { Length: > 0 } configuredPath
      ? configuredPath
      : "orchestration_journal.json";
  ```

  ```go Go
  import (
  	"bytes"
  	"cmp"
  	"context"
  	"crypto/sha256"
  	"encoding/hex"
  	"encoding/json"
  	"errors"
  	"fmt"
  	"log"
  	"os"
  	"os/exec"
  	"path/filepath"
  	"strings"
  	"sync"
  	"time"

  	"github.com/anthropics/anthropic-sdk-go"
  )

  var client = anthropic.NewClient()

  const (
  	modelID = anthropic.ModelClaudeOpus5_5
  	effort  = anthropic.OutputConfigEffortXhigh

  	systemPrompt = "You are a helpful general-purpose agent. Answer the user's request directly."

  	requestTimeoutSeconds  = 600
  	bashTimeoutSeconds     = 60
  	toolResultMaxChars     = 8000
  	maxConcurrent          = 10
  	turnsBetweenRefreshers = 10
  )

  var (
  	docTestMode      = os.Getenv("DOC_TEST_MODE") != ""
  	maxTotalSubtasks = ifTest(2, 200)
  	maxSubagentTurns = ifTest(1, 15)
  	maxMainTurns     = ifTest(1, 30)
  	journalPath      = cmp.Or(os.Getenv("ORCH_JOURNAL"), "orchestration_journal.json")
  )

  func ifTest(test, normal int) int {
  	if docTestMode {
  		return test
  	}
  	return normal
  }
  ```

  ```java Java
  // java.base (java.util, java.nio.file, java.util.concurrent, ...) se importa
  // implícitamente en un archivo fuente compacto, así que solo el SDK y Jackson necesitan imports.
  import com.anthropic.client.AnthropicClient;
  import com.anthropic.client.okhttp.AnthropicOkHttpClient;
  import com.anthropic.core.JsonArray;
  import com.anthropic.core.JsonBoolean;
  import com.anthropic.core.JsonObject;
  import com.anthropic.core.JsonString;
  import com.anthropic.core.JsonValue;
  import com.anthropic.core.RequestOptions;
  import com.anthropic.helpers.MessageAccumulator;
  import com.anthropic.models.messages.ContentBlockParam;
  import com.anthropic.models.messages.Message;
  import com.anthropic.models.messages.MessageCreateParams;
  import com.anthropic.models.messages.MessageParam;
  import com.anthropic.models.messages.Model;
  import com.anthropic.models.messages.OutputConfig;
  import com.anthropic.models.messages.StopReason;
  import com.anthropic.models.messages.TextBlock;
  import com.anthropic.models.messages.Tool;
  import com.anthropic.models.messages.ToolBash20250124;
  import com.anthropic.models.messages.ToolResultBlockParam;
  import com.anthropic.models.messages.ToolUseBlock;
  import com.fasterxml.jackson.core.JsonProcessingException;
  import com.fasterxml.jackson.core.type.TypeReference;
  import com.fasterxml.jackson.databind.JsonNode;
  import com.fasterxml.jackson.databind.ObjectMapper;

  AnthropicClient client = AnthropicOkHttpClient.fromEnv();

  static final Model MODEL = Model.CLAUDE_OPUS_5_5;
  static final boolean DOC_TEST_MODE =
          !Objects.requireNonNullElse(System.getenv("DOC_TEST_MODE"), "").isEmpty();
  static final OutputConfig.Effort EFFORT = OutputConfig.Effort.XHIGH;

  static final String SYSTEM_PROMPT =
          "You are a helpful general-purpose agent. Answer the user's request directly.";

  static final int REQUEST_TIMEOUT_SECONDS = 600;
  static final RequestOptions REQUEST_OPTIONS =
          RequestOptions.builder().timeout(Duration.ofSeconds(REQUEST_TIMEOUT_SECONDS)).build();
  static final int BASH_TIMEOUT_SECONDS = 60;
  static final int TOOL_RESULT_MAX_CHARS = 8000;
  static final int MAX_CONCURRENT = 10;
  static final int MAX_TOTAL_SUBTASKS = DOC_TEST_MODE ? 2 : 200;
  static final int MAX_SUBAGENT_TURNS = DOC_TEST_MODE ? 1 : 15;
  static final int MAX_MAIN_TURNS = DOC_TEST_MODE ? 1 : 30;
  static final int TURNS_BETWEEN_REFRESHERS = 10;
  static final Path JOURNAL_PATH = Path.of(Optional.ofNullable(System.getenv("ORCH_JOURNAL"))
          .filter(path -> !path.isEmpty())
          .orElse("orchestration_journal.json"));
  ```

  ```php PHP
  use Anthropic\Client;
  use Anthropic\Messages\ToolUseBlock;

  $client = new Client();

  const MODEL = 'claude-opus-5-5';
  define('DOC_TEST_MODE', (string) getenv('DOC_TEST_MODE') !== '');
  const EFFORT = 'xhigh';

  const SYSTEM_PROMPT = 'You are a helpful general-purpose agent. Answer the user\'s request directly.';

  const REQUEST_TIMEOUT_SECONDS = 600;
  const BASH_TIMEOUT_SECONDS = 60;
  const TOOL_RESULT_MAX_CHARS = 8000;
  const MAX_CONCURRENT = 10;
  define('MAX_TOTAL_SUBTASKS', DOC_TEST_MODE ? 2 : 200);
  define('MAX_SUBAGENT_TURNS', DOC_TEST_MODE ? 1 : 15);
  define('MAX_MAIN_TURNS', DOC_TEST_MODE ? 1 : 30);
  const TURNS_BETWEEN_REFRESHERS = 10;
  define('JOURNAL_PATH', getenv('ORCH_JOURNAL') ?: 'orchestration_journal.json');
  ```

  ```ruby Ruby
  require "anthropic"
  require "digest"
  require "fileutils"
  require "json"
  require "open3"
  require "tmpdir"

  CLIENT = Anthropic::Client.new

  MODEL = "claude-opus-5-5"
  EFFORT = :xhigh

  SYSTEM_PROMPT = "You are a helpful general-purpose agent. Answer the user's request directly."

  REQUEST_TIMEOUT_SECONDS = 600
  BASH_TIMEOUT_SECONDS = 60
  TOOL_RESULT_MAX_CHARS = 8000
  MAX_CONCURRENT = 10
  DOC_TEST_MODE = !ENV["DOC_TEST_MODE"].to_s.empty?
  MAX_TOTAL_SUBTASKS = DOC_TEST_MODE ? 2 : 200
  MAX_SUBAGENT_TURNS = DOC_TEST_MODE ? 1 : 15
  MAX_MAIN_TURNS = DOC_TEST_MODE ? 1 : 30
  TURNS_BETWEEN_REFRESHERS = 10
  JOURNAL_PATH = ENV["ORCH_JOURNAL"].to_s.empty? ? "orchestration_journal.json" : ENV["ORCH_JOURNAL"]
  ```
</CodeGroup>

## Define los recordatorios de modo

Los recordatorios son breves a propósito. Cambian el estado del modo y remiten a la descripción de la herramienta, donde se encuentran las instrucciones extensas. El texto completo se envía una vez cuando el modo se activa, el recordatorio breve se vuelve a enviar solo después de varios turnos del usuario, y el aviso de salida se envía una vez cuando el modo se desactiva.

<CodeGroup exclude="shell:cURL, shell:CLI">
  ```python Python
  MODE_ENTER = (
      "Orchestration mode is on: optimize for the most exhaustive, correct answer rather than "
      "the fastest one. Use the Workflow tool on every substantive task, sized to the problem's "
      "natural decomposition rather than the maximum the tool allows. See the Workflow tool's "
      "description for standing consent, granularity guidance, and quality patterns. Work solo "
      "only on conversational or trivial turns."
  )
  MODE_REFRESH = "Orchestration mode is still on. Use the Workflow tool; see its standing consent section."
  MODE_EXIT = (
      "Orchestration mode is off. The Workflow tool's standard opt-in rule applies again."
  )
  ```

  ```typescript TypeScript
  const MODE_ENTER =
    "Orchestration mode is on: optimize for the most exhaustive, correct answer rather than " +
    "the fastest one. Use the Workflow tool on every substantive task, sized to the problem's " +
    "natural decomposition rather than the maximum the tool allows. See the Workflow tool's " +
    "description for standing consent, granularity guidance, and quality patterns. Work solo " +
    "only on conversational or trivial turns.";
  const MODE_REFRESH =
    "Orchestration mode is still on. Use the Workflow tool; see its standing consent section.";
  const MODE_EXIT =
    "Orchestration mode is off. The Workflow tool's standard opt-in rule applies again.";
  ```

  ```csharp C#
  const string modeEnter =
      "Orchestration mode is on: optimize for the most exhaustive, correct answer rather than "
      + "the fastest one. Use the Workflow tool on every substantive task, sized to the problem's "
      + "natural decomposition rather than the maximum the tool allows. See the Workflow tool's "
      + "description for standing consent, granularity guidance, and quality patterns. Work solo "
      + "only on conversational or trivial turns.";
  const string modeRefresh =
      "Orchestration mode is still on. Use the Workflow tool; see its standing consent section.";
  const string modeExit =
      "Orchestration mode is off. The Workflow tool's standard opt-in rule applies again.";
  ```

  ```go Go
  const (
  	modeEnter = "Orchestration mode is on: optimize for the most exhaustive, correct answer rather than " +
  		"the fastest one. Use the Workflow tool on every substantive task, sized to the problem's " +
  		"natural decomposition rather than the maximum the tool allows. See the Workflow tool's " +
  		"description for standing consent, granularity guidance, and quality patterns. Work solo " +
  		"only on conversational or trivial turns."
  	modeRefresh = "Orchestration mode is still on. Use the Workflow tool; see its standing consent section."
  	modeExit    = "Orchestration mode is off. The Workflow tool's standard opt-in rule applies again."
  )
  ```

  ```java Java
  static final String MODE_ENTER = """
          Orchestration mode is on: optimize for the most exhaustive, correct answer rather than \
          the fastest one. Use the Workflow tool on every substantive task, sized to the problem's \
          natural decomposition rather than the maximum the tool allows. See the Workflow tool's \
          description for standing consent, granularity guidance, and quality patterns. Work solo \
          only on conversational or trivial turns.""";
  static final String MODE_REFRESH =
          "Orchestration mode is still on. Use the Workflow tool; see its standing consent section.";
  static final String MODE_EXIT =
          "Orchestration mode is off. The Workflow tool's standard opt-in rule applies again.";
  ```

  ```php PHP
  const MODE_ENTER =
      'Orchestration mode is on: optimize for the most exhaustive, correct answer rather than '
      . 'the fastest one. Use the Workflow tool on every substantive task, sized to the problem\'s '
      . 'natural decomposition rather than the maximum the tool allows. See the Workflow tool\'s '
      . 'description for standing consent, granularity guidance, and quality patterns. Work solo '
      . 'only on conversational or trivial turns.';
  const MODE_REFRESH =
      'Orchestration mode is still on. Use the Workflow tool; see its standing consent section.';
  const MODE_EXIT =
      'Orchestration mode is off. The Workflow tool\'s standard opt-in rule applies again.';
  ```

  ```ruby Ruby
  MODE_ENTER =
    "Orchestration mode is on: optimize for the most exhaustive, correct answer rather than " \
    "the fastest one. Use the Workflow tool on every substantive task, sized to the problem's " \
    "natural decomposition rather than the maximum the tool allows. See the Workflow tool's " \
    "description for standing consent, granularity guidance, and quality patterns. Work solo " \
    "only on conversational or trivial turns."
  MODE_REFRESH =
    "Orchestration mode is still on. Use the Workflow tool; see its standing consent section."
  MODE_EXIT =
    "Orchestration mode is off. The Workflow tool's standard opt-in rule applies again."
  ```
</CodeGroup>

## Otorga consentimiento permanente en la descripción de la herramienta

La herramienta Workflow contiene el verdadero contrato de comportamiento: la regla de activación explícita, el consentimiento permanente que se aplica mientras el modo está activado, la orientación sobre la granularidad para dimensionar la distribución en paralelo y los patrones de calidad a los que puede recurrir el modelo (una ola de verificación, un crítico de completitud y una secuenciación en varias fases). Los subagentes también reciben una herramienta `report_findings` para que sus resultados regresen como JSON estructurado en lugar de prosa. La herramienta bash es la herramienta `bash_20250124` definida por Anthropic, ejecutada localmente.

<CodeGroup exclude="shell:cURL, shell:CLI">
  ```python Python
  WORKFLOW_TOOL = {
      "name": "Workflow",
      "description": (
          "Orchestrate a multiagent workflow: split a large task into independent subtasks "
          "and run them as parallel agents, then collect their results.\n\n"
          "Opt-in: only use this tool when the user explicitly asks for a workflow, or when a "
          "system message confirms that orchestration mode is on.\n\n"
          "Quality patterns: adversarial verification (a second wave of agents checks the first "
          "wave's findings against the source), a completeness critic (one agent hunts for what "
          "the others missed), and multiphase sequencing (understand, design, implement, and "
          "review as separate workflow calls, reading results between phases). A useful default "
          "is hybrid: scout inline first to discover the work-list, then fan out over it.\n\n"
          "Granularity: scope each subtask to a distinct concern, component, or question rather "
          "than per line or per file section. Scale the count to what the user asked for: a "
          "focused review of a module of a few hundred lines rarely needs more than about ten "
          "subtasks; a broad audit of a large codebase can justify more.\n\n"
          "Standing consent: while a system message confirms orchestration mode is on, that "
          "opt-in is standing. Author and run a workflow for every substantive task by default, "
          "and lean toward verifying findings adversarially. Work solo only on conversational "
          "turns or trivial mechanical edits. When a system message says the mode is off, "
          "revert to the opt-in rule above."
      ),
      "input_schema": {
          "type": "object",
          "properties": {
              "subtasks": {
                  "type": "array",
                  "items": {"type": "string"},
                  "description": "Independent subtask prompts to run as parallel agents",
              }
          },
          "required": ["subtasks"],
      },
  }

  BASH_TOOL = {"type": "bash_20250124", "name": "bash"}

  REPORT_TOOL = {
      "name": "report_findings",
      "description": (
          "Report the final findings for your subtask. Call this exactly once, when you are "
          "done investigating; it ends your task."
      ),
      "input_schema": {
          "type": "object",
          "properties": {
              "summary": {
                  "type": "string",
                  "description": "Two or three sentences of synthesis",
              },
              "findings": {
                  "type": "array",
                  "items": {
                      "type": "object",
                      "properties": {
                          "claim": {
                              "type": "string",
                              "description": "The finding, one sentence",
                          },
                          "evidence": {
                              "type": "string",
                              "description": "How it was verified (file, line, or command output)",
                          },
                          "severity": {
                              "type": "string",
                              "enum": ["high", "medium", "low", "info"],
                          },
                      },
                      "required": ["claim", "evidence", "severity"],
                  },
              },
          },
          "required": ["summary", "findings"],
      },
  }
  ```

  ```typescript TypeScript
  const WORKFLOW_TOOL: Anthropic.Tool = {
    name: "Workflow",
    description: [
      "Orchestrate a multiagent workflow: split a large task into independent subtasks " +
        "and run them as parallel agents, then collect their results.",
      "Opt-in: only use this tool when the user explicitly asks for a workflow, or when a " +
        "system message confirms that orchestration mode is on.",
      "Quality patterns: adversarial verification (a second wave of agents checks the first " +
        "wave's findings against the source), a completeness critic (one agent hunts for what " +
        "the others missed), and multiphase sequencing (understand, design, implement, and " +
        "review as separate workflow calls, reading results between phases). A useful default " +
        "is hybrid: scout inline first to discover the work-list, then fan out over it.",
      "Granularity: scope each subtask to a distinct concern, component, or question rather " +
        "than per line or per file section. Scale the count to what the user asked for: a " +
        "focused review of a module of a few hundred lines rarely needs more than about ten " +
        "subtasks; a broad audit of a large codebase can justify more.",
      "Standing consent: while a system message confirms orchestration mode is on, that " +
        "opt-in is standing. Author and run a workflow for every substantive task by default, " +
        "and lean toward verifying findings adversarially. Work solo only on conversational " +
        "turns or trivial mechanical edits. When a system message says the mode is off, " +
        "revert to the opt-in rule above."
    ].join("\n\n"),
    input_schema: {
      type: "object",
      properties: {
        subtasks: {
          type: "array",
          items: { type: "string" },
          description: "Independent subtask prompts to run as parallel agents"
        }
      },
      required: ["subtasks"]
    }
  };

  const BASH_TOOL: Anthropic.ToolBash20250124 = { type: "bash_20250124", name: "bash" };

  const REPORT_TOOL: Anthropic.Tool = {
    name: "report_findings",
    description:
      "Report the final findings for your subtask. Call this exactly once, when you are " +
      "done investigating; it ends your task.",
    input_schema: {
      type: "object",
      properties: {
        summary: { type: "string", description: "Two or three sentences of synthesis" },
        findings: {
          type: "array",
          items: {
            type: "object",
            properties: {
              claim: { type: "string", description: "The finding, one sentence" },
              evidence: {
                type: "string",
                description: "How it was verified (file, line, or command output)"
              },
              severity: { type: "string", enum: ["high", "medium", "low", "info"] }
            },
            required: ["claim", "evidence", "severity"]
          }
        }
      },
      required: ["summary", "findings"]
    }
  };
  ```

  ```csharp C#
  Tool workflowTool = new()
  {
      Name = "Workflow",
      Description = """
          Orchestrate a multiagent workflow: split a large task into independent subtasks and run them as parallel agents, then collect their results.

          Opt-in: only use this tool when the user explicitly asks for a workflow, or when a system message confirms that orchestration mode is on.

          Quality patterns: adversarial verification (a second wave of agents checks the first wave's findings against the source), a completeness critic (one agent hunts for what the others missed), and multiphase sequencing (understand, design, implement, and review as separate workflow calls, reading results between phases). A useful default is hybrid: scout inline first to discover the work-list, then fan out over it.

          Granularity: scope each subtask to a distinct concern, component, or question rather than per line or per file section. Scale the count to what the user asked for: a focused review of a module of a few hundred lines rarely needs more than about ten subtasks; a broad audit of a large codebase can justify more.

          Standing consent: while a system message confirms orchestration mode is on, that opt-in is standing. Author and run a workflow for every substantive task by default, and lean toward verifying findings adversarially. Work solo only on conversational turns or trivial mechanical edits. When a system message says the mode is off, revert to the opt-in rule above.
          """,
      InputSchema = new()
      {
          Properties = new Dictionary<string, JsonElement>
          {
              ["subtasks"] = JsonSerializer.SerializeToElement(new
              {
                  type = "array",
                  items = new { type = "string" },
                  description = "Independent subtask prompts to run as parallel agents",
              }),
          },
          Required = ["subtasks"],
      },
  };

  ToolBash20250124 bashTool = new();

  Tool reportTool = new()
  {
      Name = "report_findings",
      Description =
          "Report the final findings for your subtask. Call this exactly once, when you are "
          + "done investigating; it ends your task.",
      InputSchema = new()
      {
          Properties = new Dictionary<string, JsonElement>
          {
              ["summary"] = JsonSerializer.SerializeToElement(new
              {
                  type = "string",
                  description = "Two or three sentences of synthesis",
              }),
              ["findings"] = JsonSerializer.SerializeToElement(new
              {
                  type = "array",
                  items = new
                  {
                      type = "object",
                      properties = new
                      {
                          claim = new { type = "string", description = "The finding, one sentence" },
                          evidence = new
                          {
                              type = "string",
                              description = "How it was verified (file, line, or command output)",
                          },
                          severity = new { type = "string", @enum = new[] { "high", "medium", "low", "info" } },
                      },
                      required = new[] { "claim", "evidence", "severity" },
                  },
              }),
          },
          Required = ["summary", "findings"],
      },
  };
  ```

  ```go Go
  var workflowTool = anthropic.ToolUnionParam{
  	OfTool: &anthropic.ToolParam{
  		Name: "Workflow",
  		Description: anthropic.String("Orchestrate a multiagent workflow: split a large task into independent subtasks " +
  			"and run them as parallel agents, then collect their results.\n\n" +
  			"Opt-in: only use this tool when the user explicitly asks for a workflow, or when a " +
  			"system message confirms that orchestration mode is on.\n\n" +
  			"Quality patterns: adversarial verification (a second wave of agents checks the first " +
  			"wave's findings against the source), a completeness critic (one agent hunts for what " +
  			"the others missed), and multiphase sequencing (understand, design, implement, and " +
  			"review as separate workflow calls, reading results between phases). A useful default " +
  			"is hybrid: scout inline first to discover the work-list, then fan out over it.\n\n" +
  			"Granularity: scope each subtask to a distinct concern, component, or question rather " +
  			"than per line or per file section. Scale the count to what the user asked for: a " +
  			"focused review of a module of a few hundred lines rarely needs more than about ten " +
  			"subtasks; a broad audit of a large codebase can justify more.\n\n" +
  			"Standing consent: while a system message confirms orchestration mode is on, that " +
  			"opt-in is standing. Author and run a workflow for every substantive task by default, " +
  			"and lean toward verifying findings adversarially. Work solo only on conversational " +
  			"turns or trivial mechanical edits. When a system message says the mode is off, " +
  			"revert to the opt-in rule above."),
  		InputSchema: anthropic.ToolInputSchemaParam{
  			Properties: map[string]any{
  				"subtasks": map[string]any{
  					"type":        "array",
  					"items":       map[string]any{"type": "string"},
  					"description": "Independent subtask prompts to run as parallel agents",
  				},
  			},
  			Required: []string{"subtasks"},
  		},
  	},
  }

  var bashTool = anthropic.ToolUnionParam{
  	OfBashTool20250124: &anthropic.ToolBash20250124Param{},
  }

  var reportTool = anthropic.ToolUnionParam{
  	OfTool: &anthropic.ToolParam{
  		Name: "report_findings",
  		Description: anthropic.String("Report the final findings for your subtask. Call this exactly once, when you are " +
  			"done investigating; it ends your task."),
  		InputSchema: anthropic.ToolInputSchemaParam{
  			Properties: map[string]any{
  				"summary": map[string]any{"type": "string", "description": "Two or three sentences of synthesis"},
  				"findings": map[string]any{
  					"type": "array",
  					"items": map[string]any{
  						"type": "object",
  						"properties": map[string]any{
  							"claim": map[string]any{"type": "string", "description": "The finding, one sentence"},
  							"evidence": map[string]any{
  								"type":        "string",
  								"description": "How it was verified (file, line, or command output)",
  							},
  							"severity": map[string]any{"type": "string", "enum": []string{"high", "medium", "low", "info"}},
  						},
  						"required": []string{"claim", "evidence", "severity"},
  					},
  				},
  			},
  			Required: []string{"summary", "findings"},
  		},
  	},
  }
  ```

  ```java Java
  static final Tool WORKFLOW_TOOL = Tool.builder()
          .name("Workflow")
          .description("""
                  Orchestrate a multiagent workflow: split a large task into independent subtasks \
                  and run them as parallel agents, then collect their results.

                  Opt-in: only use this tool when the user explicitly asks for a workflow, or when a \
                  system message confirms that orchestration mode is on.

                  Quality patterns: adversarial verification (a second wave of agents checks the first \
                  wave's findings against the source), a completeness critic (one agent hunts for what \
                  the others missed), and multiphase sequencing (understand, design, implement, and \
                  review as separate workflow calls, reading results between phases). A useful default \
                  is hybrid: scout inline first to discover the work-list, then fan out over it.

                  Granularity: scope each subtask to a distinct concern, component, or question rather \
                  than per line or per file section. Scale the count to what the user asked for: a \
                  focused review of a module of a few hundred lines rarely needs more than about ten \
                  subtasks; a broad audit of a large codebase can justify more.

                  Standing consent: while a system message confirms orchestration mode is on, that \
                  opt-in is standing. Author and run a workflow for every substantive task by default, \
                  and lean toward verifying findings adversarially. Work solo only on conversational \
                  turns or trivial mechanical edits. When a system message says the mode is off, \
                  revert to the opt-in rule above.""")
          .inputSchema(Tool.InputSchema.builder()
                  .properties(Tool.InputSchema.Properties.builder()
                          .putAdditionalProperty("subtasks", JsonValue.from(Map.of(
                                  "type", "array",
                                  "items", Map.of("type", "string"),
                                  "description", "Independent subtask prompts to run as parallel agents")))
                          .build())
                  .putAdditionalProperty("required", JsonValue.from(List.of("subtasks")))
                  .build())
          .build();

  static final ToolBash20250124 BASH_TOOL = ToolBash20250124.builder().build();

  static final Tool REPORT_TOOL = Tool.builder()
          .name("report_findings")
          .description("""
                  Report the final findings for your subtask. Call this exactly once, when you are \
                  done investigating; it ends your task.""")
          .inputSchema(Tool.InputSchema.builder()
                  .properties(Tool.InputSchema.Properties.builder()
                          .putAdditionalProperty("summary", JsonValue.from(Map.of(
                                  "type", "string",
                                  "description", "Two or three sentences of synthesis")))
                          .putAdditionalProperty("findings", JsonValue.from(Map.of(
                                  "type", "array",
                                  "items", Map.of(
                                          "type", "object",
                                          "properties", Map.of(
                                                  "claim", Map.of(
                                                          "type", "string",
                                                          "description", "The finding, one sentence"),
                                                  "evidence", Map.of(
                                                          "type", "string",
                                                          "description", "How it was verified (file, line, or command output)"),
                                                  "severity", Map.of(
                                                          "type", "string",
                                                          "enum", List.of("high", "medium", "low", "info"))),
                                          "required", List.of("claim", "evidence", "severity")))))
                          .build())
                  .putAdditionalProperty("required", JsonValue.from(List.of("summary", "findings")))
                  .build())
          .build();
  ```

  ```php PHP
  const WORKFLOW_TOOL = [
      'name' => 'Workflow',
      'description' =>
          'Orchestrate a multiagent workflow: split a large task into independent subtasks '
          . "and run them as parallel agents, then collect their results.\n\n"
          . 'Opt-in: only use this tool when the user explicitly asks for a workflow, or when a '
          . "system message confirms that orchestration mode is on.\n\n"
          . 'Quality patterns: adversarial verification (a second wave of agents checks the first '
          . 'wave\'s findings against the source), a completeness critic (one agent hunts for what '
          . 'the others missed), and multiphase sequencing (understand, design, implement, and '
          . 'review as separate workflow calls, reading results between phases). A useful default '
          . "is hybrid: scout inline first to discover the work-list, then fan out over it.\n\n"
          . 'Granularity: scope each subtask to a distinct concern, component, or question rather '
          . 'than per line or per file section. Scale the count to what the user asked for: a '
          . 'focused review of a module of a few hundred lines rarely needs more than about ten '
          . "subtasks; a broad audit of a large codebase can justify more.\n\n"
          . 'Standing consent: while a system message confirms orchestration mode is on, that '
          . 'opt-in is standing. Author and run a workflow for every substantive task by default, '
          . 'and lean toward verifying findings adversarially. Work solo only on conversational '
          . 'turns or trivial mechanical edits. When a system message says the mode is off, '
          . 'revert to the opt-in rule above.',
      'input_schema' => [
          'type' => 'object',
          'properties' => [
              'subtasks' => [
                  'type' => 'array',
                  'items' => ['type' => 'string'],
                  'description' => 'Independent subtask prompts to run as parallel agents',
              ],
          ],
          'required' => ['subtasks'],
      ],
  ];

  const BASH_TOOL = ['type' => 'bash_20250124', 'name' => 'bash'];

  const REPORT_TOOL = [
      'name' => 'report_findings',
      'description' =>
          'Report the final findings for your subtask. Call this exactly once, when you are '
          . 'done investigating; it ends your task.',
      'input_schema' => [
          'type' => 'object',
          'properties' => [
              'summary' => ['type' => 'string', 'description' => 'Two or three sentences of synthesis'],
              'findings' => [
                  'type' => 'array',
                  'items' => [
                      'type' => 'object',
                      'properties' => [
                          'claim' => ['type' => 'string', 'description' => 'The finding, one sentence'],
                          'evidence' => [
                              'type' => 'string',
                              'description' => 'How it was verified (file, line, or command output)',
                          ],
                          'severity' => ['type' => 'string', 'enum' => ['high', 'medium', 'low', 'info']],
                      ],
                      'required' => ['claim', 'evidence', 'severity'],
                  ],
              ],
          ],
          'required' => ['summary', 'findings'],
      ],
  ];
  ```

  ```ruby Ruby
  WORKFLOW_TOOL = {
    name: "Workflow",
    description:
      "Orchestrate a multiagent workflow: split a large task into independent subtasks " \
      "and run them as parallel agents, then collect their results.\n\n" \
      "Opt-in: only use this tool when the user explicitly asks for a workflow, or when a " \
      "system message confirms that orchestration mode is on.\n\n" \
      "Quality patterns: adversarial verification (a second wave of agents checks the first " \
      "wave's findings against the source), a completeness critic (one agent hunts for what " \
      "the others missed), and multiphase sequencing (understand, design, implement, and " \
      "review as separate workflow calls, reading results between phases). A useful default " \
      "is hybrid: scout inline first to discover the work-list, then fan out over it.\n\n" \
      "Granularity: scope each subtask to a distinct concern, component, or question rather " \
      "than per line or per file section. Scale the count to what the user asked for: a " \
      "focused review of a module of a few hundred lines rarely needs more than about ten " \
      "subtasks; a broad audit of a large codebase can justify more.\n\n" \
      "Standing consent: while a system message confirms orchestration mode is on, that " \
      "opt-in is standing. Author and run a workflow for every substantive task by default, " \
      "and lean toward verifying findings adversarially. Work solo only on conversational " \
      "turns or trivial mechanical edits. When a system message says the mode is off, " \
      "revert to the opt-in rule above.",
    input_schema: {
      type: "object",
      properties: {
        subtasks: {
          type: "array",
          items: {type: "string"},
          description: "Independent subtask prompts to run as parallel agents"
        }
      },
      required: ["subtasks"]
    }
  }.freeze

  BASH_TOOL = {type: "bash_20250124", name: "bash"}.freeze

  REPORT_TOOL = {
    name: "report_findings",
    description:
      "Report the final findings for your subtask. Call this exactly once, when you are " \
      "done investigating; it ends your task.",
    input_schema: {
      type: "object",
      properties: {
        summary: {type: "string", description: "Two or three sentences of synthesis"},
        findings: {
          type: "array",
          items: {
            type: "object",
            properties: {
              claim: {type: "string", description: "The finding, one sentence"},
              evidence: {
                type: "string",
                description: "How it was verified (file, line, or command output)"
              },
              severity: {type: "string", enum: ["high", "medium", "low", "info"]}
            },
            required: ["claim", "evidence", "severity"]
          }
        }
      },
      required: ["summary", "findings"]
    }
  }.freeze
  ```
</CodeGroup>

## Ejecuta la herramienta bash localmente

El controlador de bash ejecuta el comando solicitado con un tiempo de espera, captura stdout y stderr combinados y trunca el resultado para que un comando descontrolado no pueda saturar la "context window" (ventana de contexto).

Los comandos se ejecutan en el directorio desde el que inicias el ejemplo, así que para trabajar sobre un proyecto debes iniciarlo allí. Cuando `DOC_TEST_MODE` está definida, el sistema de pruebas le asigna a bash un pequeño directorio de prueba desechable que se elimina al salir.

Aquí no hay ningún entorno aislado: el comando se ejecuta con los permisos del proceso que inició el ejemplo.

Por claridad, este ejemplo ejecuta cada llamada en una subshell nueva en lugar de mantener la sesión persistente que describe el contrato de `bash_20250124`. Un agente de producción debería respaldar la herramienta con una shell de larga duración, de modo que el directorio de trabajo, el entorno y la acción `restart` se comporten según lo documentado.

<CodeGroup exclude="shell:cURL, shell:CLI">
  ```python Python
  # Ejecuta bash donde se lanzó el ejemplo. En DOC_TEST_MODE el arnés de docs
  # lo apunta en su lugar a un directorio de fixtures desechable, eliminado al salir.
  if DOC_TEST_MODE:
      WORK_DIR = tempfile.mkdtemp(prefix="orchestration-")
      atexit.register(shutil.rmtree, WORK_DIR, ignore_errors=True)
      with open(os.path.join(WORK_DIR, "sample.py"), "w") as fixture:
          fixture.write(
              "def fib(n):\n"
              "    return n if n < 2 else fib(n - 1) + fib(n - 2)\n\n"
              "print(fib(10))\n"
          )
  else:
      WORK_DIR = os.getcwd()


  def run_bash(command: str) -> tuple[str, bool]:
      """Run a shell command and return (output, is_error). No sandbox: example code only."""
      print(f"[bash] {command}", file=sys.stderr)
      try:
          proc = subprocess.run(
              ["bash", "-c", command],
              cwd=WORK_DIR,
              capture_output=True,
              text=True,
              errors="replace",
              timeout=BASH_TIMEOUT_SECONDS,
          )
      except subprocess.TimeoutExpired:
          return f"command timed out after {BASH_TIMEOUT_SECONDS}s", True
      output = (proc.stdout + proc.stderr).strip() or "(no output)"
      if len(output) > TOOL_RESULT_MAX_CHARS:
          output = (
              output[:TOOL_RESULT_MAX_CHARS]
              + f"\n(truncated at {TOOL_RESULT_MAX_CHARS} chars)"
          )
      if proc.returncode != 0:
          output = f"(exit code {proc.returncode})\n{output}"
      return output, proc.returncode != 0


  def handle_bash_block(block) -> tuple[str, bool]:
      if block.input.get("restart") is True:
          return "Shell restarted.", False
      command = block.input.get("command")
      if not isinstance(command, str) or not command:
          return "bash error: no command was provided.", True
      return run_bash(command)
  ```

  ```typescript TypeScript
  type ToolOutcome = { output: string; isError: boolean };

  // Ejecuta bash donde se lanzó el ejemplo. En DOC_TEST_MODE el arnés de docs
  // lo apunta a un directorio de fixture desechable, que se elimina al terminar el programa.
  await using fixtureDir = DOC_TEST_MODE
    ? await mkdtempDisposable(join(tmpdir(), "orchestration-"))
    : null;
  const WORK_DIR = fixtureDir?.path ?? process.cwd();
  if (fixtureDir) {
    await writeFile(
      join(WORK_DIR, "sample.py"),
      `def fib(n):
      return n if n < 2 else fib(n - 1) + fib(n - 2)

  print(fib(10))
  `
    );
  }

  // Ejecuta un comando de shell y devuelve su salida. Sin sandbox: solo código de ejemplo.
  function runBash(command: string): Promise<ToolOutcome> {
    console.error(`[bash] ${command}`);
    const options = {
      shell: "/bin/bash",
      cwd: WORK_DIR,
      timeout: BASH_TIMEOUT_SECONDS * 1000,
      maxBuffer: 16 * 1024 * 1024
    };
    return new Promise((resolve) => {
      exec(command, options, (error, stdout, stderr) => {
        if (error?.killed) {
          resolve({ output: `command timed out after ${BASH_TIMEOUT_SECONDS}s`, isError: true });
          return;
        }
        const exitCode = error === null ? 0 : typeof error.code === "number" ? error.code : 1;
        let output = (stdout + stderr).trim() || "(no output)";
        const codePoints = [...output];
        if (codePoints.length > TOOL_RESULT_MAX_CHARS) {
          const kept = codePoints.slice(0, TOOL_RESULT_MAX_CHARS).join("");
          output = `${kept}\n(truncated at ${TOOL_RESULT_MAX_CHARS} chars)`;
        }
        if (exitCode !== 0) {
          output = `(exit code ${exitCode})\n${output}`;
        }
        resolve({ output, isError: exitCode !== 0 });
      });
    });
  }

  async function handleBashBlock(block: Anthropic.ToolUseBlock): Promise<ToolOutcome> {
    const { command, restart } = block.input as { command?: string; restart?: boolean };
    if (restart === true) {
      return { output: "Shell restarted.", isError: false };
    }
    if (!command) {
      return { output: "bash error: no command was provided.", isError: true };
    }
    return runBash(command);
  }
  ```

  ```csharp C#
  // Ejecuta bash donde se lanzó el ejemplo. En DOC_TEST_MODE el arnés de docs
  // lo apunta a un directorio fixture desechable, eliminado al salir.
  var workDir = Environment.CurrentDirectory;
  if (docTestMode)
  {
      workDir = Directory.CreateTempSubdirectory("orchestration-").FullName;
      await File.WriteAllTextAsync(Path.Combine(workDir, "sample.py"), """
          def fib(n):
              return n if n < 2 else fib(n - 1) + fib(n - 2)

          print(fib(10))

          """);
      var fixtureDir = workDir;
      AppDomain.CurrentDomain.ProcessExit += (_, _) =>
      {
          try { Directory.Delete(fixtureDir, recursive: true); }
          catch { /* Best-effort cleanup; the OS tmp sweeper handles leftovers. */ }
      };
  }

  // Ejecuta un comando de shell y devuelve su salida más un flag de error. Sin sandbox: solo código de ejemplo.
  async Task<(string Output, bool IsError)> RunBash(string command)
  {
      Console.Error.WriteLine($"[bash] {command}");
      using var process = Process.Start(new ProcessStartInfo("bash")
      {
          ArgumentList = { "-c", command },
          WorkingDirectory = workDir,
          RedirectStandardOutput = true,
          RedirectStandardError = true,
      });
      if (process is null)
      {
          return ("bash error: the shell process failed to start.", true);
      }
      var stdoutTask = process.StandardOutput.ReadToEndAsync();
      var stderrTask = process.StandardError.ReadToEndAsync();
      using var timeout = new CancellationTokenSource(TimeSpan.FromSeconds(bashTimeoutSeconds));
      try
      {
          await process.WaitForExitAsync(timeout.Token);
      }
      catch (OperationCanceledException)
      {
          process.Kill(entireProcessTree: true);
          // Deja que las tareas lectoras terminen antes de que se libere el proceso.
          try
          {
              await Task.WhenAll(stdoutTask, stderrTask);
          }
          catch
          {
              // La salida se descarta al agotarse el tiempo, así que los fallos de lectura también se ignoran.
          }
          return ($"command timed out after {bashTimeoutSeconds}s", true);
      }
      var output = (await stdoutTask + await stderrTask).Trim();
      if (output.Length == 0)
      {
          output = "(no output)";
      }
      if (output.Length > toolResultMaxChars)
      {
          output = output[..toolResultMaxChars] + $"\n(truncated at {toolResultMaxChars} chars)";
      }
      if (process.ExitCode != 0)
      {
          output = $"(exit code {process.ExitCode})\n{output}";
      }
      return (output, process.ExitCode != 0);
  }

  // Ejecuta una llamada a la herramienta bash solicitada por el modelo.
  async Task<(string Output, bool IsError)> HandleBashBlock(ToolUseBlock block)
  {
      if (block.Input.GetValueOrDefault("restart").ValueKind is JsonValueKind.True)
      {
          return ("Shell restarted.", false);
      }
      return block.Input.GetValueOrDefault("command") is { ValueKind: JsonValueKind.String } command
          && command.GetString() is { Length: > 0 } commandText
          ? await RunBash(commandText)
          : ("bash error: no command was provided.", true);
  }
  ```

  ```go Go
  // Ejecuta bash donde se lanzó el ejemplo. En DOC_TEST_MODE el arnés de docs
  // lo apunta a un directorio de fixtures desechable, eliminado al salir.
  var workDir = func() string {
  	if !docTestMode {
  		dir, err := os.Getwd()
  		if err != nil {
  			log.Fatal(err)
  		}
  		return dir
  	}
  	dir, err := os.MkdirTemp("", "orchestration-")
  	if err != nil {
  		log.Fatal(err)
  	}
  	const fixture = `def fib(n):
      return n if n < 2 else fib(n - 1) + fib(n - 2)

  print(fib(10))
  `
  	if err := os.WriteFile(filepath.Join(dir, "sample.py"), []byte(fixture), 0o644); err != nil {
  		log.Fatal(err)
  	}
  	return dir
  }()

  // runBash ejecuta un comando de shell y devuelve su salida más un indicador de error.
  // Sin sandbox: solo código de ejemplo.
  func runBash(ctx context.Context, command string) (string, bool) {
  	fmt.Fprintf(os.Stderr, "[bash] %s\n", command)
  	ctx, cancel := context.WithTimeout(ctx, bashTimeoutSeconds*time.Second)
  	defer cancel()
  	cmd := exec.CommandContext(ctx, "bash", "-c", command)
  	cmd.Dir = workDir
  	combined, err := cmd.CombinedOutput()
  	if errors.Is(ctx.Err(), context.DeadlineExceeded) {
  		return fmt.Sprintf("command timed out after %ds", bashTimeoutSeconds), true
  	}
  	output := strings.TrimSpace(string(combined))
  	if output == "" {
  		output = "(no output)"
  	}
  	if runes := []rune(output); len(runes) > toolResultMaxChars {
  		output = string(runes[:toolResultMaxChars]) + fmt.Sprintf("\n(truncated at %d chars)", toolResultMaxChars)
  	}
  	if err == nil {
  		return output, false
  	}
  	if exitErr, ok := errors.AsType[*exec.ExitError](err); ok {
  		return fmt.Sprintf("(exit code %d)\n%s", exitErr.ExitCode(), output), true
  	}
  	return fmt.Sprintf("(%s)\n%s", err, output), true
  }

  // handleBashBlock ejecuta una llamada a la herramienta bash solicitada por el modelo.
  func handleBashBlock(ctx context.Context, block anthropic.ToolUseBlock) (string, bool) {
  	var input struct {
  		Command string `json:"command"`
  		Restart bool   `json:"restart"`
  	}
  	if err := json.Unmarshal(block.Input, &input); err != nil {
  		return fmt.Sprintf("bash error: could not parse input: %s", err), true
  	}
  	if input.Restart {
  		return "Shell restarted.", false
  	}
  	if input.Command == "" {
  		return "bash error: no command was provided.", true
  	}
  	return runBash(ctx, input.Command)
  }
  ```

  ```java Java
  record ToolOutput(String output, boolean isError) {}

  // Ejecuta bash donde se lanzó el ejemplo. En DOC_TEST_MODE el arnés de docs
  // lo apunta a un directorio de fixtures desechable, eliminado al salir.
  static final Path WORK_DIR = createWorkDir();

  static Path createWorkDir() {
      if (!DOC_TEST_MODE) {
          return Path.of(System.getProperty("user.dir"));
      }
      try {
          var dir = Files.createTempDirectory("orchestration-");
          Files.writeString(dir.resolve("sample.py"), """
                  def fib(n):
                      return n if n < 2 else fib(n - 1) + fib(n - 2)

                  print(fib(10))
                  """);
          Runtime.getRuntime().addShutdownHook(new Thread(() -> {
              // Limpieza de mejor esfuerzo; el limpiador de tmp del SO se encarga de los restos.
              try (var paths = Files.walk(dir)) {
                  paths.sorted(Comparator.reverseOrder()).forEach(path -> {
                      try { Files.deleteIfExists(path); } catch (IOException _) {}
                  });
              } catch (IOException _) {}
          }));
          return dir;
      } catch (IOException error) {
          throw new UncheckedIOException(error);
      }
  }

  // Ejecuta un comando de shell y devuelve su salida más un indicador de error. Sin sandbox: solo código de ejemplo.
  ToolOutput runBash(String command) throws InterruptedException {
      System.err.println("[bash] " + command);
      Process process;
      try {
          process = new ProcessBuilder("bash", "-c", command)
                  .directory(WORK_DIR.toFile())
                  .redirectErrorStream(true)
                  .start();
      } catch (IOException error) {
          return new ToolOutput("(" + error + ")", true);
      }
      // Drena stdout en otro hilo para que un pipe lleno no bloquee la espera de timeout de abajo.
      CompletableFuture<String> outputReader = CompletableFuture.supplyAsync(() -> {
          try (var stdout = process.getInputStream()) {
              return new String(stdout.readAllBytes(), StandardCharsets.UTF_8);
          } catch (IOException _) {
              return "";
          }
      });
      if (!process.waitFor(BASH_TIMEOUT_SECONDS, TimeUnit.SECONDS)) {
          process.destroyForcibly();
          outputReader.cancel(true);
          return new ToolOutput("command timed out after " + BASH_TIMEOUT_SECONDS + "s", true);
      }
      String output = outputReader.join().strip();
      if (output.isEmpty()) {
          output = "(no output)";
      } else if (output.length() > TOOL_RESULT_MAX_CHARS) {
          output = output.substring(0, TOOL_RESULT_MAX_CHARS)
                  + "\n(truncated at " + TOOL_RESULT_MAX_CHARS + " chars)";
      }
      int exitCode = process.exitValue();
      return exitCode == 0
              ? new ToolOutput(output, false)
              : new ToolOutput("(exit code " + exitCode + ")\n" + output, true);
  }

  // Los campos de entrada de una llamada a herramienta, o un mapa vacío si el modelo envió algo que no es un objeto.
  Map<String, JsonValue> toolInput(ToolUseBlock toolUse) {
      return toolUse._input() instanceof JsonObject object ? object.values() : Map.of();
  }

  // Ejecuta una llamada a la herramienta bash solicitada por el modelo.
  ToolOutput handleBashBlock(ToolUseBlock toolUse) throws InterruptedException {
      Map<String, JsonValue> input = toolInput(toolUse);
      if (input.get("restart") instanceof JsonBoolean restart && restart.value()) {
          return new ToolOutput("Shell restarted.", false);
      }
      String command = input.get("command") instanceof JsonString text ? text.value() : "";
      if (command.isEmpty()) {
          return new ToolOutput("bash error: no command was provided.", true);
      }
      return runBash(command);
  }
  ```

  ```php PHP
  // Ejecuta bash donde se lanzó el ejemplo. En DOC_TEST_MODE, el arnés de docs
  // lo apunta en su lugar a un directorio de fixtures desechable, que se elimina al salir.
  if (DOC_TEST_MODE) {
      $workDir = sys_get_temp_dir() . '/orchestration-' . bin2hex(random_bytes(8));
      if (!mkdir($workDir, 0700)) {
          throw new \RuntimeException("could not create working directory {$workDir}");
      }
      file_put_contents(
          $workDir . '/sample.py',
          "def fib(n):\n"
          . "    return n if n < 2 else fib(n - 1) + fib(n - 2)\n\n"
          . "print(fib(10))\n",
      );
      register_shutdown_function(function () use ($workDir): void {
          foreach (glob($workDir . '/*') ?: [] as $entry) {
              @unlink($entry);
          }
          @rmdir($workDir);
      });
  } else {
      $workDir = getcwd() ?: '.';
  }
  define('WORK_DIR', $workDir);

  /**
   * Run a shell command and return [output, isError]. The coreutils timeout command
   * enforces the time limit. No sandbox: example code only.
   */
  function runBash(string $command): array
  {
      fwrite(STDERR, "[bash] {$command}\n");
      // Requiere 'timeout' de GNU coreutils. En macOS: brew install coreutils, o reemplázalo con gtimeout.
      exec(
          'cd ' . escapeshellarg(WORK_DIR) . ' && timeout ' . BASH_TIMEOUT_SECONDS
              . ' bash -c ' . escapeshellarg($command) . ' 2>&1',
          $outputLines,
          $exitCode,
      );
      if ($exitCode === 124) {
          return ['command timed out after ' . BASH_TIMEOUT_SECONDS . 's', true];
      }
      $output = trim(implode("\n", $outputLines));
      if ($output === '') {
          $output = '(no output)';
      }
      if (mb_strlen($output) > TOOL_RESULT_MAX_CHARS) {
          $output = mb_substr($output, 0, TOOL_RESULT_MAX_CHARS)
              . "\n(truncated at " . TOOL_RESULT_MAX_CHARS . ' chars)';
      }
      if ($exitCode !== 0) {
          $output = "(exit code {$exitCode})\n{$output}";
      }
      return [$output, $exitCode !== 0];
  }

  /** Execute one bash tool call requested by the model. */
  function handleBashBlock(ToolUseBlock $block): array
  {
      if (($block->input['restart'] ?? null) === true) {
          return ['Shell restarted.', false];
      }
      $command = $block->input['command'] ?? '';
      if (!is_string($command) || $command === '') {
          return ['bash error: no command was provided.', true];
      }
      return runBash($command);
  }
  ```

  ```ruby Ruby
  # Ejecuta bash donde se lanzó el ejemplo. En DOC_TEST_MODE el arnés de docs
  # lo apunta en su lugar a un directorio fixture desechable, eliminado al salir.
  WORK_DIR =
    if DOC_TEST_MODE
      Dir.mktmpdir("orchestration-").tap do |dir|
        File.write(File.join(dir, "sample.py"), <<~PYTHON)
          def fib(n):
              return n if n < 2 else fib(n - 1) + fib(n - 2)

          print(fib(10))
        PYTHON
        at_exit { FileUtils.remove_entry(dir, true) }
      end
    else
      Dir.pwd
    end

  # La entrada de la herramienta llega como un Hash o como una cadena JSON sin procesar desde el acumulador
  # de streaming; normaliza cualquiera de las dos formas a un Hash con claves de cadena.
  def parse_tool_input(raw)
    return raw.transform_keys(&:to_s) if raw.is_a?(Hash)
    parsed = JSON.parse(raw.to_s) rescue nil
    parsed.is_a?(Hash) ? parsed : {}
  end

  # Ejecuta un comando de shell y devuelve [output, is_error]. Sin sandbox: solo código de ejemplo.
  def run_bash(command)
    warn "[bash] #{command}"
    begin
      stdin, stdout_and_stderr, wait_thr = Open3.popen2e("bash", "-c", command, pgroup: true, chdir: WORK_DIR)
      stdin.close
      reader = Thread.new { stdout_and_stderr.read.scrub }
      # Aplica el límite de tiempo con un plazo de reloj monotónico para que un comando que expira
      # se termine en lugar de quedar ejecutándose en segundo plano.
      deadline = Process.clock_gettime(Process::CLOCK_MONOTONIC) + BASH_TIMEOUT_SECONDS
      until wait_thr.join(0.1)
        next if Process.clock_gettime(Process::CLOCK_MONOTONIC) < deadline

        begin
          Process.kill("-TERM", wait_thr.pid)
        rescue Errno::ESRCH
        end
        unless wait_thr.join(2)
          begin
            Process.kill("-KILL", wait_thr.pid)
          rescue Errno::ESRCH
          end
        end
        wait_thr.join(5)
        reader.join(1) || reader.kill
        stdout_and_stderr.close rescue nil
        return ["command timed out after #{BASH_TIMEOUT_SECONDS}s", true]
      end
      status = wait_thr.value
      output = reader.value.strip
      stdout_and_stderr.close
      output = "(no output)" if output.empty?
      if output.length > TOOL_RESULT_MAX_CHARS
        output = "#{output[0, TOOL_RESULT_MAX_CHARS]}\n(truncated at #{TOOL_RESULT_MAX_CHARS} chars)"
      end
      output = "(exit code #{status.exitstatus})\n#{output}" unless status.success?
      [output, !status.success?]
    rescue Errno::ENOENT => e
      return ["bash error: #{e.message}", true]
    end
  end

  # Ejecuta una llamada a la herramienta bash solicitada por el modelo.
  def handle_bash_block(block)
    input = parse_tool_input(block.input)
    return ["Shell restarted.", false] if input["restart"] == true

    command = input["command"]
    return ["bash error: no command was provided.", true] unless command.is_a?(String) && !command.empty?

    run_bash(command)
  end

  # Convierte el contenido de la respuesta en parámetros con forma de solicitud. El acumulador de streaming
  # devuelve la entrada de tool_use como una cadena JSON sin procesar e incluye campos exclusivos de la respuesta,
  # así que reestructura cada bloque al esquema de solicitud antes de devolverlo.
  def assistant_content_param(content)
    content.map do |block|
      case block
      when Anthropic::Models::ToolUseBlock
        input = parse_tool_input(block.input)
        {type: "tool_use", id: block.id, name: block.name, input: input}
      when Anthropic::Models::TextBlock
        {type: "text", text: block.text}
      when Anthropic::Models::ThinkingBlock
        {type: "thinking", thinking: block.thinking, signature: block.signature}
      when Anthropic::Models::RedactedThinkingBlock then {type: "redacted_thinking", data: block.data}
      else
        block.to_h
      end
    end
  end
  ```
</CodeGroup>

## Ejecuta un subagente

Cada subtarea del flujo de trabajo se convierte en su propio pequeño bucle de agente con la herramienta bash, y se ejecuta con el mismo esfuerzo que el bucle principal. Un tiempo de espera por solicitud limita cada llamada a la API, de modo que una conexión perdida solo afecta a un subagente en lugar de detener toda la ejecución.

<CodeGroup exclude="shell:cURL, shell:CLI">
  ```python Python
  def run_subagent(model: str, prompt: str) -> str:
      """One subagent: a small nested agent loop with the bash tool plus report_findings.
      Subagents inherit the main loop's effort level."""
      subagent_system = (
          "You are one agent in a larger parallel fan-out, assigned a single subtask. "
          "Investigate it directly, using bash to check facts rather than guessing, and finish "
          "by calling report_findings exactly once. Return findings, not narration."
      )
      messages = [{"role": "user", "content": prompt}]
      for _ in range(MAX_SUBAGENT_TURNS):
          with client.messages.stream(
              model=model,
              max_tokens=64000,
              system=subagent_system,
              output_config={"effort": EFFORT},
              tools=[BASH_TOOL, REPORT_TOOL],
              messages=messages,
              timeout=REQUEST_TIMEOUT_SECONDS,
          ) as stream:
              response = stream.get_final_message()
          messages.append({"role": "assistant", "content": response.content})
          if response.stop_reason == "pause_turn":
              continue
          if response.stop_reason != "tool_use":
              text = "".join(
                  block.text for block in response.content if block.type == "text"
              )
              if response.stop_reason == "max_tokens":
                  text += "\n\n(warning: subagent response was truncated at max_tokens)"
              return text
          tool_results = []
          report = None
          for block in response.content:
              if block.type != "tool_use":
                  continue
              match block.name:
                  case "report_findings":
                      report = json.dumps(block.input, indent=2)
                      output, is_error = "Findings recorded.", False
                  case "bash":
                      output, is_error = handle_bash_block(block)
                  case _:
                      output, is_error = f"unknown tool: {block.name}", True
              tool_results.append(
                  {
                      "type": "tool_result",
                      "tool_use_id": block.id,
                      "content": output,
                      "is_error": is_error,
                  }
              )
          if report is not None:
              return report
          messages.append({"role": "user", "content": tool_results})
      return "(subagent hit the turn limit before finishing)"
  ```

  ```typescript TypeScript
  // Un subagente: un pequeño bucle de agente anidado con la herramienta bash más report_findings.
  // Los subagentes heredan el nivel de esfuerzo del bucle principal.
  async function runSubagent(model: string, prompt: string): Promise<string> {
    const subagentSystem =
      "You are one agent in a larger parallel fan-out, assigned a single subtask. " +
      "Investigate it directly, using bash to check facts rather than guessing, and finish " +
      "by calling report_findings exactly once. Return findings, not narration.";
    const messages: Anthropic.MessageParam[] = [{ role: "user", content: prompt }];
    for (let turn = 0; turn < MAX_SUBAGENT_TURNS; turn++) {
      const response = await client.messages
        .stream(
          {
            model,
            max_tokens: 64000,
            system: subagentSystem,
            output_config: { effort: EFFORT },
            tools: [BASH_TOOL, REPORT_TOOL],
            messages
          },
          { signal: AbortSignal.timeout(REQUEST_TIMEOUT_SECONDS * 1000) }
        )
        .finalMessage();
      messages.push({ role: "assistant", content: response.content });
      if (response.stop_reason === "pause_turn") {
        continue;
      }
      if (response.stop_reason !== "tool_use") {
        let text = response.content
          .flatMap((block) => (block.type === "text" ? [block.text] : []))
          .join("");
        if (response.stop_reason === "max_tokens") {
          text += "\n\n(warning: subagent response was truncated at max_tokens)";
        }
        return text;
      }
      const toolResults: Anthropic.ToolResultBlockParam[] = [];
      let report: string | null = null;
      for (const block of response.content) {
        if (block.type !== "tool_use") {
          continue;
        }
        let outcome: ToolOutcome;
        switch (block.name) {
          case "report_findings":
            report = JSON.stringify(block.input, null, 2);
            outcome = { output: "Findings recorded.", isError: false };
            break;
          case "bash":
            outcome = await handleBashBlock(block);
            break;
          default:
            outcome = { output: `unknown tool: ${block.name}`, isError: true };
        }
        toolResults.push({
          type: "tool_result",
          tool_use_id: block.id,
          content: outcome.output,
          is_error: outcome.isError
        });
      }
      if (report !== null) {
        return report;
      }
      messages.push({ role: "user", content: toolResults });
    }
    return "(subagent hit the turn limit before finishing)";
  }
  ```

  ```csharp C#
  // Un subagente: un pequeño bucle de agente anidado con la herramienta bash más report_findings.
  // Los subagentes heredan el nivel de esfuerzo del bucle principal.
  JsonSerializerOptions indented = new() { WriteIndented = true };

  async Task<string> RunSubagent(string prompt)
  {
      const string subagentSystem =
          "You are one agent in a larger parallel fan-out, assigned a single subtask. "
          + "Investigate it directly, using bash to check facts rather than guessing, and finish "
          + "by calling report_findings exactly once. Return findings, not narration.";
      List<MessageParam> messages = [new() { Role = Role.User, Content = prompt }];
      for (var turn = 0; turn < maxSubagentTurns; turn++)
      {
          using var deadline = new CancellationTokenSource(TimeSpan.FromSeconds(requestTimeoutSeconds));
          var response = await client.Messages.CreateStreaming(new MessageCreateParams
          {
              Model = model,
              MaxTokens = 64000,
              System = subagentSystem,
              OutputConfig = new OutputConfig { Effort = effort },
              Tools = [bashTool, reportTool],
              Messages = messages,
          }, cancellationToken: deadline.Token).Aggregate();
          messages.Add(new()
          {
              Role = Role.Assistant,
              Content = response.Content.Select(block => new ContentBlockParam(block.Json)).ToList(),
          });
          if (response.StopReason == StopReason.PauseTurn)
          {
              continue;
          }
          if (response.StopReason != StopReason.ToolUse)
          {
              var text = string.Concat(
                  response.Content.Select(block => block.TryPickText(out var textBlock) ? textBlock.Text : ""));
              if (response.StopReason == StopReason.MaxTokens)
              {
                  text += "\n\n(warning: subagent response was truncated at max_tokens)";
              }
              return text;
          }
          List<ContentBlockParam> toolResults = [];
          string? report = null;
          foreach (var block in response.Content)
          {
              if (!block.TryPickToolUse(out var toolUse))
              {
                  continue;
              }
              if (toolUse.Name == "report_findings")
              {
                  report = JsonSerializer.Serialize(toolUse.Input, indented);
              }
              var (output, isError) = toolUse.Name switch
              {
                  "report_findings" => ("Findings recorded.", false),
                  "bash" => await HandleBashBlock(toolUse),
                  _ => ($"unknown tool: {toolUse.Name}", true),
              };
              toolResults.Add(new ToolResultBlockParam(toolUse.ID) { Content = output, IsError = isError });
          }
          if (report is not null)
          {
              return report;
          }
          messages.Add(new() { Role = Role.User, Content = toolResults });
      }
      return "(subagent hit the turn limit before finishing)";
  }
  ```

  ```go Go
  // streamMessage envía una solicitud y acumula los eventos de streaming en el
  // mensaje final.
  func streamMessage(ctx context.Context, params anthropic.MessageNewParams) (anthropic.Message, error) {
  	ctx, cancel := context.WithTimeout(ctx, requestTimeoutSeconds*time.Second)
  	defer cancel()
  	stream := client.Messages.NewStreaming(ctx, params)
  	defer stream.Close()
  	var message anthropic.Message
  	for stream.Next() {
  		if err := message.Accumulate(stream.Current()); err != nil {
  			return message, err
  		}
  	}
  	return message, stream.Err()
  }

  // textOf une los bloques de texto de una respuesta.
  func textOf(message anthropic.Message) string {
  	var text strings.Builder
  	for _, block := range message.Content {
  		if textBlock, ok := block.AsAny().(anthropic.TextBlock); ok {
  			text.WriteString(textBlock.Text)
  		}
  	}
  	return text.String()
  }

  // runSubagent ejecuta un subagente: un pequeño bucle de agente anidado con la herramienta bash más
  // report_findings. Los subagentes heredan el nivel de esfuerzo del bucle principal.
  func runSubagent(ctx context.Context, model string, prompt string) (string, error) {
  	const subagentSystem = "You are one agent in a larger parallel fan-out, assigned a single subtask. " +
  		"Investigate it directly, using bash to check facts rather than guessing, and finish " +
  		"by calling report_findings exactly once. Return findings, not narration."
  	messages := []anthropic.MessageParam{anthropic.NewUserMessage(anthropic.NewTextBlock(prompt))}
  	for range maxSubagentTurns {
  		response, err := streamMessage(ctx, anthropic.MessageNewParams{
  			Model:        model,
  			MaxTokens:    64000,
  			System:       []anthropic.TextBlockParam{{Text: subagentSystem}},
  			OutputConfig: anthropic.OutputConfigParam{Effort: effort},
  			Tools:        []anthropic.ToolUnionParam{bashTool, reportTool},
  			Messages:     messages,
  		})
  		if err != nil {
  			return "", err
  		}
  		messages = append(messages, response.ToParam())

  		switch response.StopReason {
  		case anthropic.StopReasonPauseTurn:
  			continue
  		case anthropic.StopReasonMaxTokens:
  			return textOf(response) + "\n\n(warning: subagent response was truncated at max_tokens)", nil
  		case anthropic.StopReasonToolUse:
  		default:
  			return textOf(response), nil
  		}

  		var toolResults []anthropic.ContentBlockParamUnion
  		var report string
  		var reportRecorded bool
  		for _, block := range response.Content {
  			toolUse, ok := block.AsAny().(anthropic.ToolUseBlock)
  			if !ok {
  				continue
  			}
  			var output string
  			var isError bool
  			switch toolUse.Name {
  			case "report_findings":
  				report = string(toolUse.Input)
  				var pretty bytes.Buffer
  				if err := json.Indent(&pretty, toolUse.Input, "", "  "); err == nil {
  					report = pretty.String()
  				}
  				reportRecorded = true
  				output = "Findings recorded."
  			case "bash":
  				output, isError = handleBashBlock(ctx, toolUse)
  			default:
  				output, isError = "unknown tool: "+toolUse.Name, true
  			}
  			toolResults = append(toolResults, anthropic.NewToolResultBlock(toolUse.ID, output, isError))
  		}
  		if reportRecorded {
  			return report, nil
  		}
  		messages = append(messages, anthropic.NewUserMessage(toolResults...))
  	}
  	return "(subagent hit the turn limit before finishing)", nil
  }
  ```

  ```java Java
  static final String SUBAGENT_SYSTEM = """
          You are one agent in a larger parallel fan-out, assigned a single subtask. \
          Investigate it directly, using bash to check facts rather than guessing, and finish \
          by calling report_findings exactly once. Return findings, not narration.""";

  // Un subagente: un pequeño bucle de agente anidado con la herramienta bash más report_findings.
  // Los subagentes heredan el nivel de esfuerzo del bucle principal.
  String runSubagent(Model model, String prompt) throws InterruptedException {
      List<MessageParam> messages = new ArrayList<>();
      messages.add(MessageParam.builder().role(MessageParam.Role.USER).content(prompt).build());
      for (int turn = 0; turn < MAX_SUBAGENT_TURNS; turn++) {
          MessageCreateParams params = MessageCreateParams.builder()
                  .model(model)
                  .maxTokens(64000L)
                  .system(SUBAGENT_SYSTEM)
                  .outputConfig(OutputConfig.builder().effort(EFFORT).build())
                  .addTool(BASH_TOOL)
                  .addTool(REPORT_TOOL)
                  .messages(messages)
                  .build();
          MessageAccumulator accumulator = MessageAccumulator.create();
          try (var stream = client.messages().createStreaming(params, REQUEST_OPTIONS)) {
              stream.stream().forEach(accumulator::accumulate);
          }
          Message response = accumulator.message();
          messages.add(response.toParam());
          StopReason stopReason = response.stopReason().orElse(null);
          if (StopReason.PAUSE_TURN.equals(stopReason)) {
              continue;
          }
          if (!StopReason.TOOL_USE.equals(stopReason)) {
              String text = response.content().stream()
                      .flatMap(block -> block.text().stream())
                      .map(TextBlock::text)
                      .collect(Collectors.joining());
              if (StopReason.MAX_TOKENS.equals(stopReason)) {
                  text += "\n\n(warning: subagent response was truncated at max_tokens)";
              }
              return text;
          }
          List<ToolUseBlock> toolUses = response.content().stream()
                  .flatMap(block -> block.toolUse().stream())
                  .toList();
          List<ContentBlockParam> toolResults = new ArrayList<>();
          String report = null;
          for (ToolUseBlock toolUse : toolUses) {
              ToolOutput result = switch (toolUse.name()) {
                  case "report_findings" -> {
                      report = toolUse._input().convert(JsonNode.class).toPrettyString();
                      yield new ToolOutput("Findings recorded.", false);
                  }
                  case "bash" -> handleBashBlock(toolUse);
                  default -> new ToolOutput("unknown tool: " + toolUse.name(), true);
              };
              toolResults.add(ContentBlockParam.ofToolResult(ToolResultBlockParam.builder()
                      .toolUseId(toolUse.id())
                      .content(result.output())
                      .isError(result.isError())
                      .build()));
          }
          if (report != null) {
              return report;
          }
          messages.add(MessageParam.builder()
                  .role(MessageParam.Role.USER)
                  .contentOfBlockParams(toolResults)
                  .build());
      }
      return "(subagent hit the turn limit before finishing)";
  }
  ```

  ```php PHP
  /**
   * Consume a message stream and assemble the final assistant turn from its events:
   * the full content-block list plus the stop reason, equivalent to what a
   * non-streaming create call returns.
   */
  function drainMessageStream(iterable $events): array
  {
      $stringValue = fn ($value) => $value instanceof \BackedEnum ? $value->value : $value;
      $blocks = [];
      $jsonBuffers = [];
      $stopReason = null;
      foreach ($events as $event) {
          $type = $stringValue($event->type);
          if ($type === 'content_block_start') {
              $blocks[$event->index] = $event->contentBlock;
              $jsonBuffers[$event->index] = '';
          } elseif ($type === 'content_block_delta') {
              $block = $blocks[$event->index];
              $delta = $event->delta;
              $deltaType = $stringValue($delta->type);
              if ($deltaType === 'text_delta') {
                  $blocks[$event->index] = $block->withText($block->text . $delta->text);
              } elseif ($deltaType === 'input_json_delta') {
                  $jsonBuffers[$event->index] .= $delta->partialJSON;
              } elseif ($deltaType === 'thinking_delta') {
                  $blocks[$event->index] = $block->withThinking($block->thinking . $delta->thinking);
              } elseif ($deltaType === 'signature_delta') {
                  $blocks[$event->index] = $block->withSignature($delta->signature);
              }
          } elseif ($type === 'message_delta') {
              $stopReason = $stringValue($event->delta->stopReason);
          }
      }
      foreach ($jsonBuffers as $index => $buffer) {
          if ($buffer !== '' && $blocks[$index] instanceof \Anthropic\Messages\ToolUseBlock) {
              $decoded = json_decode($buffer, true);
              $blocks[$index] = $blocks[$index]->withInput(is_array($decoded) ? $decoded : []);
          }
      }
      return [array_values($blocks), $stopReason];
  }

  /**
   * One subagent: a small nested agent loop with the bash tool plus report_findings.
   * Subagents inherit the main loop's effort level.
   */
  function runSubagent(Client $client, string $model, string $prompt): string
  {
      $subagentSystem =
          'You are one agent in a larger parallel fan-out, assigned a single subtask. '
          . 'Investigate it directly, using bash to check facts rather than guessing, and finish '
          . 'by calling report_findings exactly once. Return findings, not narration.';
      $messages = [['role' => 'user', 'content' => $prompt]];
      for ($turn = 0; $turn < MAX_SUBAGENT_TURNS; $turn++) {
          $stream = $client->messages->createStream(
              model: $model,
              maxTokens: 64000,
              system: $subagentSystem,
              outputConfig: ['effort' => EFFORT],
              tools: [BASH_TOOL, REPORT_TOOL],
              messages: $messages,
              requestOptions: ['timeout' => REQUEST_TIMEOUT_SECONDS],
          );
          [$content, $stopReason] = drainMessageStream($stream);
          $messages[] = ['role' => 'assistant', 'content' => $content];
          if ($stopReason === 'pause_turn') {
              continue;
          }
          if ($stopReason !== 'tool_use') {
              $text = '';
              foreach ($content as $block) {
                  if ($block instanceof \Anthropic\Messages\TextBlock) {
                      $text .= $block->text;
                  }
              }
              if ($stopReason === 'max_tokens') {
                  $text .= "\n\n(warning: subagent response was truncated at max_tokens)";
              }
              return $text;
          }
          $report = null;
          $toolResults = [];
          foreach ($content as $block) {
              if (!$block instanceof \Anthropic\Messages\ToolUseBlock) {
                  continue;
              }
              if ($block->name === 'report_findings') {
                  $report = json_encode($block->input, JSON_PRETTY_PRINT);
                  $output = 'Findings recorded.';
                  $isError = false;
              } elseif ($block->name === 'bash') {
                  [$output, $isError] = handleBashBlock($block);
              } else {
                  $output = "unknown tool: {$block->name}";
                  $isError = true;
              }
              $toolResults[] = [
                  'type' => 'tool_result',
                  'tool_use_id' => $block->id,
                  'content' => $output,
                  'is_error' => $isError,
              ];
          }
          if ($report !== null) {
              return $report;
          }
          $messages[] = ['role' => 'user', 'content' => $toolResults];
      }
      return '(subagent hit the turn limit before finishing)';
  }
  ```

  ```ruby Ruby
  # Un subagente: un pequeño bucle de agente anidado con la herramienta bash más report_findings.
  # Los subagentes heredan el nivel de esfuerzo del bucle principal.
  def run_subagent(model, prompt)
    subagent_system =
      "You are one agent in a larger parallel fan-out, assigned a single subtask. " \
      "Investigate it directly, using bash to check facts rather than guessing, and finish " \
      "by calling report_findings exactly once. Return findings, not narration."
    messages = [{role: "user", content: prompt}]
    MAX_SUBAGENT_TURNS.times do
      stream = CLIENT.messages.stream(
        model: model,
        max_tokens: 64_000,
        system_: subagent_system,
        output_config: {effort: EFFORT},
        tools: [BASH_TOOL, REPORT_TOOL],
        messages: messages,
        request_options: {timeout: REQUEST_TIMEOUT_SECONDS}
      )
      response = stream.accumulated_message
      messages << {role: "assistant", content: assistant_content_param(response.content)}
      next if response.stop_reason == :pause_turn

      unless response.stop_reason == :tool_use
        text = response.content.select { |block| block.type == :text }.map(&:text).join
        text += "\n\n(warning: subagent response was truncated at max_tokens)" if response.stop_reason == :max_tokens
        return text
      end

      report = nil
      tool_results = []
      response.content.each do |block|
        next unless block.type == :tool_use

        input = parse_tool_input(block.input)
        case block.name
        when "report_findings"
          report = JSON.pretty_generate(input)
          output, is_error = "Findings recorded.", false
        when "bash"
          output, is_error = handle_bash_block(block)
        else
          output, is_error = "unknown tool: #{block.name}", true
        end
        tool_results << {
          type: "tool_result",
          tool_use_id: block.id,
          content: output,
          is_error: is_error
        }
      end
      return report unless report.nil?

      messages << {role: "user", content: tool_results}
    end
    "(subagent hit the turn limit before finishing)"
  end
  ```
</CodeGroup>

## Registra los resultados en un diario para que las nuevas ejecuciones se reanuden

Reiniciar desde cero una distribución en paralelo que genera decenas de subagentes es costoso. Un pequeño "journal" (diario) direccionado por contenido hace que el proceso sea idempotente: antes de despachar un subagente, busca el SHA-256 de su prompt en un archivo JSON local y, si existe un resultado registrado, lo devuelve.

Si interrumpes la ejecución y la vuelves a lanzar, solo se recalculan las subtareas que nunca terminaron. El diario elimina duplicados entre ejecuciones, no dentro de una misma ola de distribución en paralelo. Para empezar de cero, elimina el archivo del diario.

<CodeGroup exclude="shell:cURL, shell:CLI">
  ```python Python
  _journal_lock = threading.Lock()


  def _load_journal() -> dict:
      try:
          with open(JOURNAL_PATH) as file:
              return json.load(file) or {}
      except (OSError, json.JSONDecodeError):
          return {}


  def journaled(prompt: str, compute) -> str:
      """Return a cached result for this exact prompt, or compute and persist it. This
      makes the fan-out resumable: interrupt the run, rerun it, and only the subtasks
      that never finished are recomputed. Delete the journal file to start fresh."""
      key = hashlib.sha256(prompt.encode()).hexdigest()
      cached = _load_journal().get(key)
      if cached is not None:
          print(f"[journal] cache hit for {key[:12]}", file=sys.stderr)
          return cached
      result = compute()
      try:
          with _journal_lock:  # fan-out writes from many threads
              journal = _load_journal()
              journal[key] = result
              temp = f"{JOURNAL_PATH}.tmp"
              with open(temp, "w") as file:
                  json.dump(journal, file)
              os.replace(temp, JOURNAL_PATH)  # atomic on POSIX and Windows
      # Una escritura fallida en el diario no debe perder el resultado calculado
      except OSError as error:
          print(f"[journal] write failed: {error}", file=sys.stderr)
      return result
  ```

  ```typescript TypeScript
  let journalWriteChain = Promise.resolve();

  async function loadJournal(): Promise<Record<string, string>> {
    try {
      return JSON.parse(await readFile(JOURNAL_PATH, "utf8")) ?? {};
    } catch (error) {
      if ((error as NodeJS.ErrnoException).code !== "ENOENT") {
        console.error(`[journal] discarding unreadable journal: ${error}`);
      }
      return {};
    }
  }

  // Devuelve un resultado en caché para este prompt exacto, o lo calcula y lo persiste. Esto
  // hace que el fan-out sea reanudable: interrumpe la ejecución, vuelve a ejecutarla, y solo las subtareas
  // que nunca terminaron se recalculan. Elimina el archivo de journal para empezar de cero.
  async function journaled(prompt: string, compute: () => Promise<string>): Promise<string> {
    const key = createHash("sha256").update(prompt).digest("hex");
    const cached = (await loadJournal())[key];
    if (cached !== undefined) {
      console.error(`[journal] cache hit for ${key.slice(0, 12)}`);
      return cached;
    }
    const result = await compute();
    // Encadena las escrituras para que los subagentes concurrentes no sobrescriban las entradas de los demás.
    // La cadena se mantiene resuelta para que una escritura fallida no contamine las siguientes.
    await (journalWriteChain = journalWriteChain
      .then(async () => {
        const journal = await loadJournal();
        journal[key] = result;
        const temp = `${JOURNAL_PATH}.tmp`;
        await writeFile(temp, JSON.stringify(journal));
        await rename(temp, JOURNAL_PATH);
      })
      .catch((error) => console.error(`[journal] write failed: ${error}`)));
    return result;
  }
  ```

  ```csharp C#
  SemaphoreSlim journalLock = new(1, 1);

  async Task<Dictionary<string, string>> LoadJournal()
  {
      try
      {
          return JsonSerializer.Deserialize<Dictionary<string, string>>(await File.ReadAllTextAsync(journalPath)) ?? [];
      }
      catch (Exception error) when (error is IOException or UnauthorizedAccessException or JsonException)
      {
          return [];
      }
  }

  // Devuelve un resultado en caché para este prompt exacto, o calcúlalo y persístelo. Esto
  // hace que la distribución sea reanudable: interrumpe la ejecución, vuelve a lanzarla, y solo
  // se recalculan las subtareas que nunca terminaron. Borra el archivo de journal para empezar de cero.
  async Task<string> Journaled(string prompt, Func<Task<string>> compute)
  {
      var key = Convert.ToHexStringLower(SHA256.HashData(Encoding.UTF8.GetBytes(prompt)));
      if ((await LoadJournal()).TryGetValue(key, out var cached))
      {
          Console.Error.WriteLine($"[journal] cache hit for {key[..12]}");
          return cached;
      }
      var result = await compute();
      await journalLock.WaitAsync(); // fan-out writes from many tasks
      try
      {
          var journal = await LoadJournal();
          journal[key] = result;
          var temp = journalPath + ".tmp";
          await File.WriteAllTextAsync(temp, JsonSerializer.Serialize(journal));
          File.Move(temp, journalPath, overwrite: true);
      }
      catch (Exception error) when (error is IOException or UnauthorizedAccessException or NotSupportedException)
      {
          // El journal es best-effort; nunca descartes un resultado ya calculado.
          Console.Error.WriteLine($"[journal] write failed: {error.Message}");
      }
      finally
      {
          journalLock.Release();
      }
      return result;
  }
  ```

  ```go Go
  var journalMutex sync.Mutex

  func loadJournal() map[string]string {
  	data, err := os.ReadFile(journalPath)
  	if err != nil {
  		return map[string]string{}
  	}
  	var journal map[string]string
  	if err := json.Unmarshal(data, &journal); err != nil || journal == nil {
  		return map[string]string{}
  	}
  	return journal
  }

  // journaled devuelve un resultado en caché para este prompt exacto, o lo calcula y persiste.
  // Esto hace que la distribución sea reanudable: interrumpe la ejecución, vuelve a ejecutarla, y solo
  // se recalculan las subtareas que nunca terminaron. Borra el archivo de journal para empezar de cero.
  func journaled(prompt string, compute func() (string, error)) (string, error) {
  	sum := sha256.Sum256([]byte(prompt))
  	key := hex.EncodeToString(sum[:])
  	if cached, ok := loadJournal()[key]; ok {
  		fmt.Fprintf(os.Stderr, "[journal] cache hit for %s\n", key[:12])
  		return cached, nil
  	}
  	result, err := compute()
  	if err != nil {
  		return "", err
  	}
  	journalMutex.Lock() // fan-out writes from many goroutines
  	defer journalMutex.Unlock()
  	journal := loadJournal()
  	journal[key] = result
  	data, _ := json.Marshal(journal)
  	temp := journalPath + ".tmp"
  	if err := os.WriteFile(temp, data, 0o644); err != nil {
  		fmt.Fprintf(os.Stderr, "[journal] write failed: %s\n", err)
  	} else if err := os.Rename(temp, journalPath); err != nil {
  		fmt.Fprintf(os.Stderr, "[journal] write failed: %s\n", err)
  		_ = os.Remove(temp)
  	}
  	return result, nil
  }
  ```

  ```java Java
  static final ObjectMapper JOURNAL_MAPPER = new ObjectMapper();
  static final ReentrantLock JOURNAL_LOCK = new ReentrantLock();

  Map<String, String> loadJournal() {
      try {
          return Objects.requireNonNullElseGet(
                  JOURNAL_MAPPER.readValue(Files.readString(JOURNAL_PATH), new TypeReference<HashMap<String, String>>() {}),
                  HashMap::new);
      } catch (IOException _) {
          return new HashMap<>();
      }
  }

  // Devuelve un resultado en caché para este prompt exacto, o calcúlalo y persístelo. Esto
  // hace la distribución reanudable: interrumpe la ejecución, reejecútala, y solo las subtareas
  // que nunca terminaron se recalculan. Elimina el archivo de journal para empezar de cero.
  String journaled(String prompt, Callable<String> compute) throws Exception {
      var digest = MessageDigest.getInstance("SHA-256").digest(prompt.getBytes(StandardCharsets.UTF_8));
      String key = HexFormat.of().formatHex(digest);
      String cached = loadJournal().get(key);
      if (cached != null) {
          System.err.println("[journal] cache hit for " + key.substring(0, 12));
          return cached;
      }
      String result = compute.call();
      JOURNAL_LOCK.lock(); // fan-out writes from many threads
      try {
          Map<String, String> journal = loadJournal();
          journal.put(key, result);
          Path temp = JOURNAL_PATH.resolveSibling(JOURNAL_PATH.getFileName() + ".tmp");
          Files.writeString(temp, JOURNAL_MAPPER.writeValueAsString(journal));
          Files.move(temp, JOURNAL_PATH, StandardCopyOption.REPLACE_EXISTING, StandardCopyOption.ATOMIC_MOVE);
      } catch (IOException error) {
          // El journal es de mejor esfuerzo; nunca descartes un resultado calculado.
          System.err.println("[journal] write failed: " + error);
      } finally {
          JOURNAL_LOCK.unlock();
      }
      return result;
  }
  ```

  ```php PHP
  function loadJournal(): array
  {
      $raw = @file_get_contents(JOURNAL_PATH);
      if ($raw === false) {
          return [];
      }
      $decoded = json_decode($raw, true);
      return is_array($decoded) ? $decoded : [];
  }

  /**
   * Return a cached result for this exact prompt, or compute and persist it. This
   * makes the fan-out resumable: interrupt the run, rerun it, and only the subtasks
   * that never finished are recomputed. Delete the journal file to start fresh.
   */
  function journaled(string $prompt, callable $compute): string
  {
      $key = hash('sha256', $prompt);
      $journal = loadJournal();
      if (array_key_exists($key, $journal)) {
          fwrite(STDERR, '[journal] cache hit for ' . substr($key, 0, 12) . "\n");
          return $journal[$key];
      }
      $result = $compute();
      $journal = loadJournal();
      $journal[$key] = $result;
      $temp = JOURNAL_PATH . '.tmp';
      $encoded = json_encode($journal, JSON_INVALID_UTF8_SUBSTITUTE);
      if ($encoded === false || @file_put_contents($temp, $encoded) === false || !@rename($temp, JOURNAL_PATH)) {
          fwrite(STDERR, '[journal] write failed: ' . (error_get_last()['message'] ?? json_last_error_msg()) . "\n");
          @unlink($temp);
      }
      return $result;
  }
  ```

  ```ruby Ruby
  JOURNAL_LOCK = Mutex.new

  def load_journal
    JSON.parse(File.read(JOURNAL_PATH)) || {}
  rescue SystemCallError, JSON::ParserError
    {}
  end

  # Devuelve un resultado en caché para este prompt exacto, o calcúlalo y persístelo. Esto
  # hace que el fan-out sea reanudable: interrumpe la ejecución, vuelve a ejecutarla y solo se recalculan
  # las subtareas que nunca terminaron. Elimina el archivo de journal para empezar de cero.
  def journaled(prompt)
    key = Digest::SHA256.hexdigest(prompt)
    cached = load_journal[key]
    unless cached.nil?
      warn "[journal] cache hit for #{key[0, 12]}"
      return cached
    end
    result = yield
    begin
      JOURNAL_LOCK.synchronize do # fan-out writes from many threads
        journal = load_journal
        journal[key] = result
        temp = "#{JOURNAL_PATH}.tmp"
        File.write(temp, JSON.generate(journal))
        File.rename(temp, JOURNAL_PATH)
      end
    rescue SystemCallError => error # the journal is best-effort; never discard a computed result
      warn "[journal] write failed: #{error}"
    end
    result
  end
  ```
</CodeGroup>

## Distribuye en paralelo y luego verifica

La distribución en paralelo acepta hasta `MAX_TOTAL_SUBTASKS` prompts y los ejecuta a través del diario con un máximo de `MAX_CONCURRENT` en curso (de forma secuencial en la versión en PHP). Además, aísla los fallos: si un subagente falla, su resultado se reduce a una cadena de error en lugar de terminar la ejecución.

Cuando termina la primera ola, una segunda ola reutiliza la misma ruta de subagentes para intentar refutar cada resultado. Cada verificador vuelve a derivar las afirmaciones a partir de la fuente y, en caso de duda, las considera refutadas. Tanto el resultado original como su veredicto se devuelven al orquestador para que pueda sopesarlos en conjunto.

<CodeGroup exclude="shell:cURL, shell:CLI">
  ```python Python
  def normalize_subtasks(raw) -> list[str]:
      """Accept the subtasks input in whatever shape the model emits: an array, the array
      JSON-encoded as a single string, or a newline-separated list."""
      if isinstance(raw, str):
          try:
              raw = json.loads(raw)
          except json.JSONDecodeError:
              raw = raw.splitlines() if "\n" in raw else [raw]
      if not isinstance(raw, list):
          return []
      return [task.strip() for task in raw if isinstance(task, str) and task.strip()]


  def verify_prompt_for(subtask: str, result: str) -> str:
      return (
          "Adversarially verify the subagent result below: try to REFUTE it. Re-derive the "
          "claims yourself with bash rather than trusting the result, and look for evidence "
          "that contradicts them. Default to refuted if uncertain. Call report_findings with "
          "summary 'refuted: <why>' or 'confirmed: <why>', citing the file:line or command "
          "output that decided it.\n\n"
          f"""Subtask: {subtask}

  Result to verify:
  {result}"""
      )


  def run_workflow(model: str, raw_subtasks) -> tuple[str, bool]:
      """Run subtasks as parallel subagents, then run a second verification wave over
      the results, and return both. MAX_TOTAL_SUBTASKS bounds how many the model can
      queue; MAX_CONCURRENT bounds how many run at once."""
      all_subtasks = normalize_subtasks(raw_subtasks)
      subtasks = all_subtasks[:MAX_TOTAL_SUBTASKS]
      dropped = len(all_subtasks) - len(subtasks)
      if not subtasks:
          return "Workflow error: no usable subtasks were provided.", True
      print(f"[workflow] fanning out {len(subtasks)} agents", file=sys.stderr)

      def run_one(prompt: str) -> str:
          try:
              return journaled(prompt, lambda: run_subagent(model, prompt))
          # Un subagente fallido no debe terminar la ejecución
          except Exception as error:
              return f"(subagent failed: {type(error).__name__}: {error})"

      with concurrent.futures.ThreadPoolExecutor(max_workers=MAX_CONCURRENT) as pool:
          results = list(pool.map(run_one, subtasks))
          print(f"[workflow] verifying {len(results)} results", file=sys.stderr)
          verify_prompts = [
              verify_prompt_for(task, result) for task, result in zip(subtasks, results)
          ]
          verdicts = list(pool.map(run_one, verify_prompts))

      joined = "\n\n".join(
          f"""[agent {index + 1}: {task}]
  {result}

  [verify {index + 1}]
  {verdict}"""
          for index, (task, result, verdict) in enumerate(
              zip(subtasks, results, verdicts)
          )
      )
      if dropped > 0:
          joined = (
              f"(note: {dropped} subtasks beyond MAX_TOTAL_SUBTASKS={MAX_TOTAL_SUBTASKS} were not "
              "run; rerun them in a follow-up Workflow call)\n\n" + joined
          )
      return joined, False
  ```

  ```typescript TypeScript
  // Acepta la entrada de subtareas en la forma que el modelo emita: un arreglo, el arreglo
  // codificado en JSON como una sola cadena, o una lista separada por saltos de línea.
  function normalizeSubtasks(raw: unknown): string[] {
    let value = raw;
    if (typeof raw === "string") {
      try {
        value = JSON.parse(raw);
      } catch {
        value = raw.includes("\n") ? raw.split("\n") : [raw];
      }
    }
    if (!Array.isArray(value)) {
      return [];
    }
    return value
      .filter((task) => typeof task === "string")
      .map((task) => task.trim())
      .filter((task) => task.length > 0);
  }

  function verifyPromptFor(subtask: string, result: string): string {
    const instructions =
      "Adversarially verify the subagent result below: try to REFUTE it. Re-derive the " +
      "claims yourself with bash rather than trusting the result, and look for evidence " +
      "that contradicts them. Default to refuted if uncertain. Call report_findings with " +
      "summary 'refuted: <why>' or 'confirmed: <why>', citing the file:line or command " +
      "output that decided it.";
    return `${instructions}

  Subtask: ${subtask}

  Result to verify:
  ${result}`;
  }

  // Map con límite de concurrencia: a lo sumo `limit` tareas en vuelo a la vez.
  async function mapWithLimit<In, Out>(
    items: readonly In[],
    limit: number,
    task: (item: In) => Promise<Out>
  ): Promise<Out[]> {
    const results = new Array<Out>(items.length);
    let cursor = 0;
    const workers = Array.from({ length: Math.min(limit, items.length) }, async () => {
      while (cursor < items.length) {
        const index = cursor++;
        results[index] = await task(items[index]);
      }
    });
    await Promise.all(workers);
    return results;
  }

  // Ejecuta las subtareas como subagentes paralelos, luego ejecuta una segunda ola de verificación sobre
  // los resultados, y devuelve ambos. MAX_TOTAL_SUBTASKS limita cuántas puede
  // encolar el modelo; MAX_CONCURRENT limita cuántas se ejecutan a la vez.
  async function runWorkflow(model: string, rawSubtasks: unknown): Promise<ToolOutcome> {
    const allSubtasks = normalizeSubtasks(rawSubtasks);
    const subtasks = allSubtasks.slice(0, MAX_TOTAL_SUBTASKS);
    const dropped = allSubtasks.length - subtasks.length;
    if (subtasks.length === 0) {
      return { output: "Workflow error: no usable subtasks were provided.", isError: true };
    }
    console.error(`[workflow] fanning out ${subtasks.length} agents`);

    const runOne = async (prompt: string): Promise<string> => {
      try {
        return await journaled(prompt, () => runSubagent(model, prompt));
      } catch (error) {
        // Límite de aislamiento: un subagente defectuoso no debería terminar la ejecución.
        const reason =
          error instanceof Error ? `${error.name}: ${error.message}` : String(error);
        return `(subagent failed: ${reason})`;
      }
    };

    const results = await mapWithLimit(subtasks, MAX_CONCURRENT, runOne);
    console.error(`[workflow] verifying ${results.length} results`);
    const verifyPrompts = subtasks.map((task, index) => verifyPromptFor(task, results[index]));
    const verdicts = await mapWithLimit(verifyPrompts, MAX_CONCURRENT, runOne);

    const report = subtasks
      .map(
        (task, index) => `[agent ${index + 1}: ${task}]
  ${results[index]}

  [verify ${index + 1}]
  ${verdicts[index]}`
      )
      .join("\n\n");
    const droppedNote =
      dropped > 0
        ? `(note: ${dropped} subtasks beyond MAX_TOTAL_SUBTASKS=${MAX_TOTAL_SUBTASKS} were ` +
          "not run; rerun them in a follow-up Workflow call)\n\n"
        : "";
    return { output: droppedNote + report, isError: false };
  }
  ```

  ```csharp C#
  // Acepta la entrada de subtareas en cualquier forma que emita el modelo: un array, el array
  // codificado en JSON como una sola cadena, o una lista separada por saltos de línea.
  List<string> NormalizeSubtasks(JsonElement raw)
  {
      IEnumerable<string?> tasks = raw.ValueKind switch
      {
          JsonValueKind.Array => raw.EnumerateArray()
              .Where(item => item.ValueKind == JsonValueKind.String)
              .Select(item => item.GetString()),
          JsonValueKind.String => ParseSubtaskString(raw.GetString()!),
          _ => [],
      };
      return [.. tasks.OfType<string>().Select(task => task.Trim()).Where(task => task.Length > 0)];
  }

  IEnumerable<string?> ParseSubtaskString(string single)
  {
      try
      {
          return JsonSerializer.Deserialize<List<string?>>(single) ?? [];
      }
      catch (JsonException)
      {
          return single.Split('\n');
      }
  }

  string VerifyPromptFor(string subtask, string result) => $"""
      Adversarially verify the subagent result below: try to REFUTE it. Re-derive the claims yourself with bash rather than trusting the result, and look for evidence that contradicts them. Default to refuted if uncertain. Call report_findings with summary 'refuted: <why>' or 'confirmed: <why>', citing the file:line or command output that decided it.

      Subtask: {subtask}

      Result to verify:
      {result}
      """;

  // Ejecuta subtareas como subagentes paralelos, luego ejecuta una segunda ola de verificación
  // sobre los resultados, y devuelve ambos. maxTotalSubtasks limita cuántas puede encolar
  // el modelo; maxConcurrent limita cuántas se ejecutan a la vez.
  async Task<(string Output, bool IsError)> RunWorkflow(JsonElement rawSubtasks)
  {
      var allSubtasks = NormalizeSubtasks(rawSubtasks);
      List<string> subtasks = [.. allSubtasks.Take(maxTotalSubtasks)];
      var dropped = allSubtasks.Count - subtasks.Count;
      if (subtasks.Count == 0)
      {
          return ("Workflow error: no usable subtasks were provided.", true);
      }
      Console.Error.WriteLine($"[workflow] fanning out {subtasks.Count} agents");

      using SemaphoreSlim gate = new(maxConcurrent);
      async Task<string> RunOne(string prompt)
      {
          await gate.WaitAsync();
          try
          {
              return await Journaled(prompt, () => RunSubagent(prompt));
          }
          catch (Exception error)
          {
              // Límite de aislamiento: un subagente defectuoso no debe terminar la ejecución.
              return $"(subagent failed: {error.GetType().Name}: {error.Message})";
          }
          finally
          {
              gate.Release();
          }
      }

      var results = await Task.WhenAll(subtasks.Select(RunOne));
      Console.Error.WriteLine($"[workflow] verifying {results.Length} results");
      var verdicts = await Task.WhenAll(subtasks.Zip(results, VerifyPromptFor).Select(RunOne));

      var joined = string.Join("\n\n", subtasks.Select((task, index) => $"""
          [agent {index + 1}: {task}]
          {results[index]}

          [verify {index + 1}]
          {verdicts[index]}
          """));
      if (dropped > 0)
      {
          joined = $"""
              (note: {dropped} subtasks beyond maxTotalSubtasks={maxTotalSubtasks} were not run; rerun them in a follow-up Workflow call)

              {joined}
              """;
      }
      return (joined, false);
  }
  ```

  ```go Go
  // normalizeSubtasks acepta la entrada de subtareas en cualquier forma que emita el modelo: un
  // array, el array codificado en JSON como una sola cadena, o una lista separada por saltos de línea.
  func normalizeSubtasks(raw json.RawMessage) []string {
  	var tasks []string
  	if err := json.Unmarshal(raw, &tasks); err != nil {
  		var single string
  		if err := json.Unmarshal(raw, &single); err != nil {
  			return nil
  		}
  		if err := json.Unmarshal([]byte(single), &tasks); err != nil {
  			tasks = strings.Split(single, "\n")
  		}
  	}
  	cleaned := make([]string, 0, len(tasks))
  	for _, task := range tasks {
  		if trimmed := strings.TrimSpace(task); trimmed != "" {
  			cleaned = append(cleaned, trimmed)
  		}
  	}
  	return cleaned
  }

  func verifyPromptFor(subtask, result string) string {
  	return "Adversarially verify the subagent result below: try to REFUTE it. Re-derive the " +
  		"claims yourself with bash rather than trusting the result, and look for evidence " +
  		"that contradicts them. Default to refuted if uncertain. Call report_findings with " +
  		"summary 'refuted: <why>' or 'confirmed: <why>', citing the file:line or command " +
  		"output that decided it.\n\n" +
  		fmt.Sprintf(`Subtask: %s

  Result to verify:
  %s`, subtask, result)
  }

  // mapWithLimit ejecuta task sobre items con como máximo limit goroutines en vuelo.
  func mapWithLimit(items []string, limit int, task func(string) string) []string {
  	results := make([]string, len(items))
  	semaphore := make(chan struct{}, limit)
  	var waitGroup sync.WaitGroup
  	for index, item := range items {
  		semaphore <- struct{}{}
  		waitGroup.Go(func() {
  			defer func() { <-semaphore }()
  			results[index] = task(item)
  		})
  	}
  	waitGroup.Wait()
  	return results
  }

  // runWorkflow ejecuta subtareas como subagentes paralelos, luego ejecuta una segunda ola de verificación
  // sobre los resultados, y devuelve ambos. maxTotalSubtasks limita cuántas puede encolar el
  // modelo; maxConcurrent limita cuántas se ejecutan a la vez.
  func runWorkflow(ctx context.Context, model string, rawSubtasks json.RawMessage) (string, bool) {
  	allSubtasks := normalizeSubtasks(rawSubtasks)
  	subtasks := allSubtasks[:min(len(allSubtasks), maxTotalSubtasks)]
  	dropped := len(allSubtasks) - len(subtasks)
  	if len(subtasks) == 0 {
  		return "Workflow error: no usable subtasks were provided.", true
  	}
  	fmt.Fprintf(os.Stderr, "[workflow] fanning out %d agents\n", len(subtasks))

  	runOne := func(prompt string) string {
  		report, err := journaled(prompt, func() (string, error) { return runSubagent(ctx, model, prompt) })
  		if err != nil {
  			// Límite de aislamiento: un subagente defectuoso no debe terminar la ejecución.
  			return fmt.Sprintf("(subagent failed: %s)", err)
  		}
  		return report
  	}

  	results := mapWithLimit(subtasks, maxConcurrent, runOne)
  	fmt.Fprintf(os.Stderr, "[workflow] verifying %d results\n", len(results))
  	verifyPrompts := make([]string, len(subtasks))
  	for index, task := range subtasks {
  		verifyPrompts[index] = verifyPromptFor(task, results[index])
  	}
  	verdicts := mapWithLimit(verifyPrompts, maxConcurrent, runOne)

  	sections := make([]string, len(subtasks))
  	for index, task := range subtasks {
  		sections[index] = fmt.Sprintf(`[agent %d: %s]
  %s

  [verify %d]
  %s`,
  			index+1, task, results[index], index+1, verdicts[index])
  	}
  	joined := strings.Join(sections, "\n\n")
  	if dropped > 0 {
  		joined = fmt.Sprintf("(note: %d subtasks beyond maxTotalSubtasks=%d were not run; "+
  			"rerun them in a follow-up Workflow call)\n\n", dropped, maxTotalSubtasks) + joined
  	}
  	return joined, false
  }
  ```

  ```java Java
  // Acepta la entrada de subtareas en cualquier forma que el modelo emita: un array, el array
  // codificado en JSON como una sola cadena, o una lista separada por saltos de línea.
  List<String> normalizeSubtasks(JsonValue rawSubtasks) {
      List<String> tasks = switch (rawSubtasks) {
          case JsonArray array -> array.values().stream()
                  .map(item -> item instanceof JsonString text ? text.value() : item.toString())
                  .toList();
          case JsonString string -> parseSubtaskString(string.value());
          default -> List.of();
      };
      return tasks.stream()
              .filter(Objects::nonNull)
              .map(String::strip)
              .filter(task -> !task.isEmpty())
              .toList();
  }

  List<String> parseSubtaskString(String encoded) {
      try {
          String[] parsed = new ObjectMapper().readValue(encoded, String[].class);
          return parsed == null ? List.of() : Arrays.asList(parsed);
      } catch (JsonProcessingException _) {
          return encoded.lines().toList();
      }
  }

  String verifyPromptFor(String subtask, String result) {
      return """
              Adversarially verify the subagent result below: try to REFUTE it. Re-derive the \
              claims yourself with bash rather than trusting the result, and look for evidence \
              that contradicts them. Default to refuted if uncertain. Call report_findings with \
              summary 'refuted: <why>' or 'confirmed: <why>', citing the file:line or command \
              output that decided it.

              Subtask: %s

              Result to verify:
              %s""".formatted(subtask, result);
  }

  // invokeAll espera a cada tarea, así que cada future ya terminó cuando se lee aquí.
  List<String> runAll(ExecutorService pool, List<String> prompts, Model model) throws InterruptedException {
      List<Callable<String>> jobs = prompts.stream()
              .<Callable<String>>map(prompt -> () -> journaled(prompt, () -> runSubagent(model, prompt)))
              .toList();
      return pool.invokeAll(jobs).stream()
              .map(future -> switch (future.state()) {
                  case SUCCESS -> future.resultNow();
                  // Límite de aislamiento: un subagente defectuoso no debe terminar la ejecución.
                  case FAILED -> "(subagent failed: " + future.exceptionNow() + ")";
                  case CANCELLED, RUNNING -> "(subagent failed: cancelled)";
              })
              .toList();
  }

  // Ejecuta subtareas como subagentes paralelos, luego ejecuta una segunda ola de verificación sobre
  // los resultados, y devuelve ambos. MAX_TOTAL_SUBTASKS limita cuántas puede
  // encolar el modelo; MAX_CONCURRENT limita cuántas se ejecutan a la vez.
  ToolOutput runWorkflow(Model model, JsonValue rawSubtasks) throws InterruptedException {
      List<String> allSubtasks = normalizeSubtasks(rawSubtasks);
      List<String> subtasks = allSubtasks.stream().limit(MAX_TOTAL_SUBTASKS).toList();
      int dropped = allSubtasks.size() - subtasks.size();
      if (subtasks.isEmpty()) {
          return new ToolOutput("Workflow error: no usable subtasks were provided.", true);
      }
      System.err.println("[workflow] fanning out " + subtasks.size() + " agents");

      List<String> results;
      List<String> verdicts;
      try (ExecutorService pool = Executors.newFixedThreadPool(MAX_CONCURRENT, Thread.ofVirtual().factory())) {
          results = runAll(pool, subtasks, model);
          System.err.println("[workflow] verifying " + results.size() + " results");
          List<String> verifyPrompts = IntStream.range(0, subtasks.size())
                  .mapToObj(index -> verifyPromptFor(subtasks.get(index), results.get(index)))
                  .toList();
          verdicts = runAll(pool, verifyPrompts, model);
      }
      String joined = IntStream.range(0, subtasks.size())
              .mapToObj(index -> """
                      [agent %d: %s]
                      %s

                      [verify %d]
                      %s""".formatted(index + 1, subtasks.get(index), results.get(index), index + 1, verdicts.get(index)))
              .collect(Collectors.joining("\n\n"));
      if (dropped > 0) {
          joined = """
                  (note: %d subtasks beyond MAX_TOTAL_SUBTASKS=%d were not run; rerun them in a follow-up Workflow call)

                  %s""".formatted(dropped, MAX_TOTAL_SUBTASKS, joined);
      }
      return new ToolOutput(joined, false);
  }
  ```

  ```php PHP
  /**
   * Accept the subtasks input in whatever shape the model emits: an array, the array
   * JSON-encoded as a single string, or a newline-separated list.
   */
  function normalizeSubtasks(mixed $raw): array
  {
      if (is_string($raw)) {
          try {
              $raw = json_decode($raw, true, flags: JSON_THROW_ON_ERROR);
          } catch (\JsonException) {
              $raw = str_contains($raw, "\n") ? explode("\n", $raw) : [$raw];
          }
      }
      if (!is_array($raw)) {
          return [];
      }
      $tasks = array_map('trim', array_filter($raw, 'is_string'));
      return array_values(array_filter($tasks, fn ($task) => $task !== ''));
  }

  function verifyPromptFor(string $subtask, string $result): string
  {
      return 'Adversarially verify the subagent result below: try to REFUTE it. Re-derive the '
          . 'claims yourself with bash rather than trusting the result, and look for evidence '
          . 'that contradicts them. Default to refuted if uncertain. Call report_findings with '
          . "summary 'refuted: <why>' or 'confirmed: <why>', citing the file:line or command "
          . "output that decided it.\n\n"
          . <<<PROMPT
              Subtask: {$subtask}

              Result to verify:
              {$result}
              PROMPT;
  }

  /**
   * Run subtasks through the journal, then run a second verification wave over the
   * results, and return both. PHP's standard runtime has no lightweight thread pool,
   * so both waves run sequentially here (MAX_CONCURRENT is unused); the SDK examples
   * in other languages fan them out in parallel.
   */
  function runWorkflow(Client $client, string $model, mixed $rawSubtasks): array
  {
      $allSubtasks = normalizeSubtasks($rawSubtasks);
      $subtasks = array_slice($allSubtasks, 0, MAX_TOTAL_SUBTASKS);
      $dropped = count($allSubtasks) - count($subtasks);
      if ($subtasks === []) {
          return ['Workflow error: no usable subtasks were provided.', true];
      }
      fwrite(STDERR, '[workflow] running ' . count($subtasks) . " agents\n");

      $runOne = function (string $prompt) use ($client, $model): string {
          try {
              return journaled($prompt, fn () => runSubagent($client, $model, $prompt));
          } catch (\Throwable $error) {
              // Límite de aislamiento: un subagente defectuoso no debería terminar la ejecución.
              return '(subagent failed: ' . $error::class . ': ' . $error->getMessage() . ')';
          }
      };

      $results = array_map($runOne, $subtasks);
      fwrite(STDERR, '[workflow] verifying ' . count($results) . " results\n");
      $verifyPrompts = array_map(verifyPromptFor(...), $subtasks, $results);
      $verdicts = array_map($runOne, $verifyPrompts);

      $sections = [];
      foreach ($subtasks as $index => $task) {
          $sections[] = '[agent ' . ($index + 1) . ": {$task}]\n{$results[$index]}"
              . "\n\n[verify " . ($index + 1) . "]\n{$verdicts[$index]}";
      }
      $joined = implode("\n\n", $sections);
      if ($dropped > 0) {
          $joined = '(note: ' . $dropped . ' subtasks beyond MAX_TOTAL_SUBTASKS=' . MAX_TOTAL_SUBTASKS
              . " were not run; rerun them in a follow-up Workflow call)\n\n" . $joined;
      }
      return [$joined, false];
  }
  ```

  ```ruby Ruby
  # Acepta la entrada de subtareas en la forma que el modelo emita: un arreglo, el arreglo
  # codificado en JSON como una sola cadena, o una lista separada por saltos de línea.
  def normalize_subtasks(raw)
    if raw.is_a?(String)
      begin
        raw = JSON.parse(raw)
      rescue JSON::ParserError
        raw = raw.include?("\n") ? raw.split("\n") : [raw]
      end
    end
    return [] unless raw.is_a?(Array)
    raw.select { |task| task.is_a?(String) }.map(&:strip).reject(&:empty?)
  end

  def verify_prompt_for(subtask, result)
    "Adversarially verify the subagent result below: try to REFUTE it. Re-derive the " \
      "claims yourself with bash rather than trusting the result, and look for evidence " \
      "that contradicts them. Default to refuted if uncertain. Call report_findings with " \
      "summary 'refuted: <why>' or 'confirmed: <why>', citing the file:line or command " \
      "output that decided it.\n\n" +
      <<~PROMPT.chomp
        Subtask: #{subtask}

        Result to verify:
        #{result}
      PROMPT
  end

  # Mapea con un límite de concurrencia: a lo sumo `limit` hilos en vuelo a la vez.
  def map_with_limit(items, limit)
    results = Array.new(items.length)
    queue = Queue.new
    items.each_with_index { |item, index| queue << [index, item] }
    workers = Array.new([limit, items.length].min) do
      Thread.new do
        until queue.empty?
          index, item = queue.pop(true) rescue break
          results[index] = yield item
        end
      end
    end
    workers.each(&:join)
    results
  end

  # Ejecuta las subtareas como subagentes paralelos, luego ejecuta una segunda ola de verificación sobre
  # los resultados y devuelve ambos. MAX_TOTAL_SUBTASKS limita cuántas puede encolar
  # el modelo; MAX_CONCURRENT limita cuántas se ejecutan a la vez.
  def run_workflow(model, raw_subtasks)
    all_subtasks = normalize_subtasks(raw_subtasks)
    subtasks = all_subtasks.first(MAX_TOTAL_SUBTASKS)
    dropped = all_subtasks.length - subtasks.length
    return ["Workflow error: no usable subtasks were provided.", true] if subtasks.empty?

    warn "[workflow] fanning out #{subtasks.length} agents"
    run_one = lambda do |prompt|
      journaled(prompt) { run_subagent(model, prompt) }
    rescue => error # isolation boundary: one bad subagent should not end the run
      "(subagent failed: #{error.class}: #{error.message})"
    end

    results = map_with_limit(subtasks, MAX_CONCURRENT, &run_one)
    warn "[workflow] verifying #{results.length} results"
    verify_prompts = subtasks.zip(results).map { |task, result| verify_prompt_for(task, result) }
    verdicts = map_with_limit(verify_prompts, MAX_CONCURRENT, &run_one)

    joined = subtasks.each_with_index.map do |task, index|
      <<~ENTRY.chomp
        [agent #{index + 1}: #{task}]
        #{results[index]}

        [verify #{index + 1}]
        #{verdicts[index]}
      ENTRY
    end.join("\n\n")
    if dropped > 0
      joined =
        "(note: #{dropped} subtasks beyond MAX_TOTAL_SUBTASKS=#{MAX_TOTAL_SUBTASKS} were not " \
        "run; rerun them in a follow-up Workflow call)\n\n#{joined}"
    end
    [joined, false]
  end
  ```
</CodeGroup>

## Activa y desactiva el modo con mensajes del sistema a mitad de conversación

El agente agrega primero el mensaje del usuario y luego los mensajes del sistema que correspondan: el aviso de salida, el texto completo del modo al activarlo o el recordatorio periódico. Colocar el mensaje del sistema después del turno del usuario tiene dos ventajas: deja intacto cada byte en caché que lo precede y cumple la regla de ubicación según la cual un mensaje del sistema debe ir después de un turno del usuario.

<CodeGroup>
  ```bash cURL
  # Un turno en modo de orquestación: el recordatorio de modo viaja en el arreglo messages como una
  # entrada {"role": "system"} colocada después del turno del usuario al que aplica. La respuesta
  # se detiene en la primera llamada a herramienta. El bucle del agente que ejecuta llamadas a herramientas y
  # distribuye subagentes se muestra en las pestañas del SDK; la descripción del Workflow está condensada
  # aquí; los ejemplos del SDK llevan el texto completo de consentimiento permanente.
  curl --fail-with-body -sS https://api.anthropic.com/v1/messages \
    -H "x-api-key: $ANTHROPIC_API_KEY" \
    -H "anthropic-version: 2023-06-01" \
    -H "content-type: application/json" \
    -d '{
      "model": "claude-opus-5-5",
      "max_tokens": 64000,
      "system": "You are a helpful general-purpose agent. Answer the request from the user directly.",
      "output_config": {"effort": "xhigh"},
      "tools": [
        {
          "name": "Workflow",
          "description": "Orchestrate a multiagent workflow: split a large task into independent subtasks and run them as parallel agents, then collect their results. Opt-in: only use this tool when the user explicitly asks for a workflow, or when a system message confirms that orchestration mode is on. Granularity: scope each subtask to a distinct concern rather than per line or file section; a review of a small module rarely needs more than about ten. Standing consent: while a system message confirms orchestration mode is on, author and run a workflow for every substantive task by default; work solo only on conversational turns or trivial mechanical edits.",
          "input_schema": {
            "type": "object",
            "properties": {
              "subtasks": {
                "type": "array",
                "items": {"type": "string"},
                "description": "Independent subtask prompts to run as parallel agents"
              }
            },
            "required": ["subtasks"]
          }
        },
        {"type": "bash_20250124", "name": "bash"}
      ],
      "messages": [
        {
          "role": "user",
          "content": "Explore the current directory, then give a thorough review: what it does, code-quality issues, and concrete improvements."
        },
        {
          "role": "system",
          "content": "Orchestration mode is on: optimize for the most exhaustive, correct answer rather than the fastest one. Use the Workflow tool on every substantive task, sized to the natural decomposition of the problem rather than the maximum the tool allows. See the description of the Workflow tool for standing consent, granularity guidance, and quality patterns. Work solo only on conversational or trivial turns."
        }
      ]
    }'
  ```

  ```bash CLI
  # Un turno en modo de orquestación: el recordatorio de modo viaja en el arreglo messages como una
  # entrada con rol system colocada después del turno de usuario al que aplica. La respuesta se detiene
  # en la primera llamada a herramienta. El bucle de agente que ejecuta llamadas a herramientas y distribuye
  # subagentes se muestra en las pestañas del SDK; la descripción del Workflow está condensada aquí,
  # los ejemplos del SDK contienen el texto completo de consentimiento permanente.
  ant messages create <<'YAML'
  model: claude-opus-5-5
  max_tokens: 64000
  system: You are a helpful general-purpose agent. Answer the user's request directly.
  output_config: {effort: xhigh}
  tools:
    - name: Workflow
      description: >-
        Orchestrate a multiagent workflow: split a large task into independent
        subtasks and run them as parallel agents, then collect their results.
        Opt-in: only use this tool when the user explicitly asks for a workflow,
        or when a system message confirms that orchestration mode is on.
        Granularity: scope each subtask to a distinct concern rather than per
        line or file section; a review of a small module rarely needs more than
        about ten. Standing consent: while a system message confirms
        orchestration mode is on, author and run a workflow for every
        substantive task by default; work solo only on conversational turns or
        trivial mechanical edits.
      input_schema:
        type: object
        properties:
          subtasks:
            type: array
            items: {type: string}
            description: Independent subtask prompts to run as parallel agents
        required: [subtasks]
    - {type: bash_20250124, name: bash}
  messages:
    - role: user
      content: >-
        Explore the current directory, then give a thorough review: what it
        does, code-quality issues, and concrete improvements.
    - role: system
      content: >-
        Orchestration mode is on: optimize for the most exhaustive, correct
        answer rather than the fastest one. Use the Workflow tool on every
        substantive task, sized to the problem's natural decomposition rather
        than the maximum the tool allows. See the Workflow tool's description
        for standing consent, granularity guidance, and quality patterns. Work
        solo only on conversational or trivial turns.
  YAML
  ```

  ```python Python
  class ModeAgent:
      """An agent loop whose orchestration mode is toggled with mid-conversation system messages."""

      def __init__(self, model: str, mode_on: bool = True):
          self.model = model
          self.mode_on = mode_on
          self.messages: list[dict] = []
          self._mode_announced = False
          self._exit_pending = False
          self._turns_since_reminder = 0

      def set_mode(self, mode_on: bool) -> None:
          """Turn the mode on or off. The notice is delivered with the next user turn."""
          if mode_on == self.mode_on:
              return
          if not mode_on:
              if self._mode_announced:
                  self._exit_pending = True
          else:
              self._exit_pending = False
          self.mode_on = mode_on

      def _due_system_messages(self) -> list[dict]:
          """System messages owed on this turn: an exit notice, the full mode text on entry,
          or a one-line refresher every TURNS_BETWEEN_REFRESHERS user turns."""
          due = []
          if self._exit_pending:
              self._exit_pending = False
              self._mode_announced = False
              due.append({"role": "system", "content": MODE_EXIT})
          if self.mode_on:
              if not self._mode_announced:
                  self._mode_announced = True
                  self._turns_since_reminder = 0
                  due.append({"role": "system", "content": MODE_ENTER})
              elif self._turns_since_reminder >= TURNS_BETWEEN_REFRESHERS:
                  self._turns_since_reminder = 0
                  due.append({"role": "system", "content": MODE_REFRESH})
          return due

      def turn(self, user_input: str) -> str:
          # Los mensajes de sistema a mitad de conversación siguen al turno del usuario al que aplican, lo que mantiene
          # intacto el prefijo en caché que los precede.
          self.messages.append({"role": "user", "content": user_input})
          self.messages.extend(self._due_system_messages())
          self._turns_since_reminder += 1

          for _ in range(MAX_MAIN_TURNS):
              with client.messages.stream(
                  model=self.model,
                  max_tokens=64000,
                  system=SYSTEM_PROMPT,  # static for the whole session
                  output_config={"effort": EFFORT},
                  tools=[WORKFLOW_TOOL, BASH_TOOL],
                  messages=self.messages,
                  timeout=REQUEST_TIMEOUT_SECONDS,
              ) as stream:
                  response = stream.get_final_message()
              self.messages.append({"role": "assistant", "content": response.content})

              if response.stop_reason == "pause_turn":
                  continue
              if response.stop_reason != "tool_use":
                  text = "".join(
                      block.text for block in response.content if block.type == "text"
                  )
                  if response.stop_reason == "max_tokens":
                      # Descarta el mensaje truncado del asistente para que los turnos posteriores no se basen en él.
                      self.messages.pop()
                      text += "\n\n(warning: response was truncated at max_tokens)"
                  return text

              tool_results = []
              for block in response.content:
                  if block.type != "tool_use":
                      continue
                  match block.name:
                      case "Workflow":
                          output, is_error = run_workflow(
                              self.model, block.input.get("subtasks", [])
                          )
                      case "bash":
                          output, is_error = handle_bash_block(block)
                      case _:
                          output, is_error = f"unknown tool: {block.name}", True
                  tool_results.append(
                      {
                          "type": "tool_result",
                          "tool_use_id": block.id,
                          "content": output,
                          "is_error": is_error,
                      }
                  )
              self.messages.append({"role": "user", "content": tool_results})
          return "(hit the main loop turn limit before finishing)"
  ```

  ```typescript TypeScript
  // Un bucle de agente cuyo modo de orquestación se alterna con mensajes del sistema a mitad de conversación.
  class ModeAgent {
    readonly #model: string;
    readonly #messages: Anthropic.MessageParam[] = [];
    #modeOn: boolean;
    #modeAnnounced = false;
    #exitPending = false;
    #turnsSinceReminder = 0;

    constructor(model: string, modeOn = true) {
      this.#model = model;
      this.#modeOn = modeOn;
    }

    // Activa o desactiva el modo. El aviso se entrega con el siguiente turno del usuario.
    setMode(modeOn: boolean): void {
      if (modeOn === this.#modeOn) {
        return;
      }
      // Solo se debe un aviso de salida si al modelo se le dijo que el modo estaba activo.
      this.#exitPending = !modeOn && this.#modeAnnounced;
      this.#modeOn = modeOn;
    }

    // Mensajes del sistema pendientes en este turno: un aviso de salida, el texto completo del modo al entrar,
    // o un recordatorio de una línea cada TURNS_BETWEEN_REFRESHERS turnos de usuario.
    #dueSystemMessages(): Anthropic.MessageParam[] {
      const due: Anthropic.MessageParam[] = [];
      if (this.#exitPending) {
        this.#exitPending = false;
        this.#modeAnnounced = false;
        due.push({ role: "system", content: MODE_EXIT });
      }
      if (this.#modeOn) {
        if (!this.#modeAnnounced) {
          this.#modeAnnounced = true;
          this.#turnsSinceReminder = 0;
          due.push({ role: "system", content: MODE_ENTER });
        } else if (this.#turnsSinceReminder >= TURNS_BETWEEN_REFRESHERS) {
          this.#turnsSinceReminder = 0;
          due.push({ role: "system", content: MODE_REFRESH });
        }
      }
      return due;
    }

    async turn(userInput: string): Promise<string> {
      // Los mensajes del sistema a mitad de conversación siguen al turno de usuario al que aplican, lo que mantiene
      // intacto el prefijo en caché que los precede.
      this.#messages.push({ role: "user", content: userInput }, ...this.#dueSystemMessages());
      this.#turnsSinceReminder += 1;

      for (let turn = 0; turn < MAX_MAIN_TURNS; turn++) {
        const response = await client.messages
          .stream(
            {
              model: this.#model,
              max_tokens: 64000,
              system: SYSTEM_PROMPT, // static for the whole session
              output_config: { effort: EFFORT },
              tools: [WORKFLOW_TOOL, BASH_TOOL],
              messages: this.#messages
            },
            { signal: AbortSignal.timeout(REQUEST_TIMEOUT_SECONDS * 1000) }
          )
          .finalMessage();
        this.#messages.push({ role: "assistant", content: response.content });

        if (response.stop_reason === "pause_turn") {
          continue;
        }
        if (response.stop_reason !== "tool_use") {
          let text = response.content
            .flatMap((block) => (block.type === "text" ? [block.text] : []))
            .join("");
          if (response.stop_reason === "max_tokens") {
            // Descarta el mensaje truncado del asistente para que los turnos posteriores no se basen en él.
            this.#messages.pop();
            text += "\n\n(warning: response was truncated at max_tokens)";
          }
          return text;
        }

        const toolResults: Anthropic.ToolResultBlockParam[] = [];
        for (const block of response.content) {
          if (block.type !== "tool_use") {
            continue;
          }
          let outcome: ToolOutcome;
          switch (block.name) {
            case "Workflow": {
              const { subtasks } = block.input as { subtasks?: unknown };
              outcome = await runWorkflow(this.#model, subtasks);
              break;
            }
            case "bash":
              outcome = await handleBashBlock(block);
              break;
            default:
              outcome = { output: `unknown tool: ${block.name}`, isError: true };
          }
          toolResults.push({
            type: "tool_result",
            tool_use_id: block.id,
            content: outcome.output,
            is_error: outcome.isError
          });
        }
        this.#messages.push({ role: "user", content: toolResults });
      }
      return "(hit the main loop turn limit before finishing)";
    }
  }
  ```

  ```csharp C#
  // Un bucle de agente cuyo modo de orquestación se alterna con mensajes de sistema a mitad de conversación.
  List<MessageParam> messages = [];
  var modeOn = true;
  var modeAnnounced = false;
  var exitPending = false;
  var turnsSinceReminder = 0;

  // Activa o desactiva el modo. El aviso se entrega con el siguiente turno del usuario.
  void SetMode(bool nextModeOn)
  {
      if (nextModeOn == modeOn)
      {
          return;
      }
      // Solo se debe un aviso de salida si se le dijo al modelo que el modo estaba activo.
      exitPending = !nextModeOn && modeAnnounced;
      modeOn = nextModeOn;
  }

  MessageParam SystemMessage(string content) => new() { Role = Role.System, Content = content };

  // Mensajes de sistema pendientes en este turno: un aviso de salida, el texto completo del modo al entrar,
  // o un recordatorio de una línea cada turnsBetweenRefreshers turnos de usuario.
  List<MessageParam> DueSystemMessages()
  {
      List<MessageParam> due = [];
      if (exitPending)
      {
          exitPending = false;
          modeAnnounced = false;
          due.Add(SystemMessage(modeExit));
      }
      if (modeOn)
      {
          if (!modeAnnounced)
          {
              modeAnnounced = true;
              turnsSinceReminder = 0;
              due.Add(SystemMessage(modeEnter));
          }
          else if (turnsSinceReminder >= turnsBetweenRefreshers)
          {
              turnsSinceReminder = 0;
              due.Add(SystemMessage(modeRefresh));
          }
      }
      return due;
  }

  // Envía un turno de usuario por el bucle, ejecutando llamadas a herramientas hasta que el modelo pare.
  async Task<string> Turn(string userInput)
  {
      // Los mensajes de sistema a mitad de conversación siguen al turno de usuario al que aplican, lo que
      // mantiene intacto el prefijo en caché que los precede.
      messages.Add(new() { Role = Role.User, Content = userInput });
      messages.AddRange(DueSystemMessages());
      turnsSinceReminder++;

      for (var turn = 0; turn < maxMainTurns; turn++)
      {
          using var deadline = new CancellationTokenSource(TimeSpan.FromSeconds(requestTimeoutSeconds));
          var response = await client.Messages.CreateStreaming(new MessageCreateParams
          {
              Model = model,
              MaxTokens = 64000,
              System = systemPrompt, // static for the whole session
              OutputConfig = new OutputConfig { Effort = effort },
              Tools = [workflowTool, bashTool],
              Messages = messages,
          }, cancellationToken: deadline.Token).Aggregate();
          messages.Add(new()
          {
              Role = Role.Assistant,
              Content = response.Content.Select(block => new ContentBlockParam(block.Json)).ToList(),
          });

          if (response.StopReason == StopReason.PauseTurn)
          {
              continue;
          }
          if (response.StopReason != StopReason.ToolUse)
          {
              var text = string.Concat(
                  response.Content.Select(block => block.TryPickText(out var textBlock) ? textBlock.Text : ""));
              if (response.StopReason == StopReason.MaxTokens)
              {
                  // Descarta el mensaje truncado del asistente para que los turnos posteriores no se basen en él.
                  messages.RemoveAt(messages.Count - 1);
                  text += "\n\n(warning: response was truncated at max_tokens)";
              }
              return text;
          }

          List<ContentBlockParam> toolResults = [];
          foreach (var block in response.Content)
          {
              if (!block.TryPickToolUse(out var toolUse))
              {
                  continue;
              }
              var (output, isError) = toolUse.Name switch
              {
                  "Workflow" => await RunWorkflow(toolUse.Input.GetValueOrDefault("subtasks")),
                  "bash" => await HandleBashBlock(toolUse),
                  _ => ($"unknown tool: {toolUse.Name}", true),
              };
              toolResults.Add(new ToolResultBlockParam(toolUse.ID) { Content = output, IsError = isError });
          }
          messages.Add(new() { Role = Role.User, Content = toolResults });
      }
      return "(hit the main loop turn limit before finishing)";
  }
  ```

  ```go Go
  // modeAgent es un bucle de agente cuyo modo de orquestación se alterna con mensajes de sistema
  // a mitad de conversación.
  type modeAgent struct {
  	model              string
  	modeOn             bool
  	messages           []anthropic.MessageParam
  	modeAnnounced      bool
  	exitPending        bool
  	turnsSinceReminder int
  }

  func newModeAgent(model string) *modeAgent {
  	return &modeAgent{model: model, modeOn: true}
  }

  // setMode activa o desactiva el modo. El aviso se entrega con el siguiente turno del usuario.
  func (agent *modeAgent) setMode(modeOn bool) {
  	if modeOn == agent.modeOn {
  		return
  	}
  	if !modeOn {
  		if agent.modeAnnounced {
  			agent.exitPending = true
  		}
  	} else {
  		agent.exitPending = false
  	}
  	agent.modeOn = modeOn
  }

  // dueSystemMessages devuelve los mensajes de sistema pendientes en este turno: un aviso de salida, el
  // texto completo del modo al entrar, o un recordatorio de una línea cada turnsBetweenRefreshers turnos.
  func (agent *modeAgent) dueSystemMessages() []anthropic.MessageParam {
  	systemMessage := func(content string) anthropic.MessageParam {
  		return anthropic.MessageParam{
  			Role:    anthropic.MessageParamRoleSystem,
  			Content: []anthropic.ContentBlockParamUnion{anthropic.NewTextBlock(content)},
  		}
  	}
  	var due []anthropic.MessageParam
  	if agent.exitPending {
  		agent.exitPending = false
  		agent.modeAnnounced = false
  		due = append(due, systemMessage(modeExit))
  	}
  	if agent.modeOn {
  		if !agent.modeAnnounced {
  			agent.modeAnnounced = true
  			agent.turnsSinceReminder = 0
  			due = append(due, systemMessage(modeEnter))
  		} else if agent.turnsSinceReminder >= turnsBetweenRefreshers {
  			agent.turnsSinceReminder = 0
  			due = append(due, systemMessage(modeRefresh))
  		}
  	}
  	return due
  }

  // turn envía un turno de usuario por el bucle, ejecutando llamadas a herramientas hasta que el modelo pare.
  func (agent *modeAgent) turn(ctx context.Context, userInput string) (string, error) {
  	// Los mensajes de sistema a mitad de conversación siguen al turno de usuario al que aplican, lo que
  	// mantiene intacto el prefijo en caché que los precede.
  	agent.messages = append(agent.messages, anthropic.NewUserMessage(anthropic.NewTextBlock(userInput)))
  	agent.messages = append(agent.messages, agent.dueSystemMessages()...)
  	agent.turnsSinceReminder++

  	for range maxMainTurns {
  		response, err := streamMessage(ctx, anthropic.MessageNewParams{
  			Model:        agent.model,
  			MaxTokens:    64000,
  			System:       []anthropic.TextBlockParam{{Text: systemPrompt}}, // static for the whole session
  			OutputConfig: anthropic.OutputConfigParam{Effort: effort},
  			Tools:        []anthropic.ToolUnionParam{workflowTool, bashTool},
  			Messages:     agent.messages,
  		})
  		if err != nil {
  			return "", err
  		}
  		agent.messages = append(agent.messages, response.ToParam())

  		switch response.StopReason {
  		case anthropic.StopReasonPauseTurn:
  			continue
  		case anthropic.StopReasonMaxTokens:
  			// Descarta el mensaje truncado del asistente para que los turnos posteriores no se basen en él.
  			agent.messages = agent.messages[:len(agent.messages)-1]
  			return textOf(response) + "\n\n(warning: response was truncated at max_tokens)", nil
  		case anthropic.StopReasonToolUse:
  		default:
  			return textOf(response), nil
  		}

  		var toolResults []anthropic.ContentBlockParamUnion
  		for _, block := range response.Content {
  			toolUse, ok := block.AsAny().(anthropic.ToolUseBlock)
  			if !ok {
  				continue
  			}
  			var output string
  			var isError bool
  			switch toolUse.Name {
  			case "Workflow":
  				var input struct {
  					Subtasks json.RawMessage `json:"subtasks"`
  				}
  				if err := json.Unmarshal(toolUse.Input, &input); err != nil {
  					output, isError = fmt.Sprintf("Workflow error: could not parse input: %s", err), true
  				} else {
  					output, isError = runWorkflow(ctx, agent.model, input.Subtasks)
  				}
  			case "bash":
  				output, isError = handleBashBlock(ctx, toolUse)
  			default:
  				output, isError = "unknown tool: "+toolUse.Name, true
  			}
  			toolResults = append(toolResults, anthropic.NewToolResultBlock(toolUse.ID, output, isError))
  		}
  		agent.messages = append(agent.messages, anthropic.NewUserMessage(toolResults...))
  	}
  	return "(hit the main loop turn limit before finishing)", nil
  }
  ```

  ```java Java
  // Un bucle de agente cuyo modo de orquestación se alterna con mensajes de sistema en mitad de la conversación.
  class ModeAgent {
      private final Model model;
      private final List<MessageParam> messages = new ArrayList<>();
      private boolean modeOn;
      private boolean modeAnnounced;
      private boolean exitPending;
      private int turnsSinceReminder;

      ModeAgent(Model model) {
          this(model, true);
      }

      ModeAgent(Model model, boolean modeOn) {
          this.model = model;
          this.modeOn = modeOn;
      }

      // Activa o desactiva el modo. El aviso se entrega con el siguiente turno del usuario:
      // solo se debe un aviso de salida si al modelo se le dijo que el modo estaba activo.
      void setMode(boolean modeOn) {
          if (modeOn == this.modeOn) {
              return;
          }
          exitPending = !modeOn && modeAnnounced;
          this.modeOn = modeOn;
      }

      // Mensajes de sistema pendientes en este turno: un aviso de salida, el texto completo del modo al entrar,
      // o un recordatorio de una línea cada TURNS_BETWEEN_REFRESHERS turnos de usuario.
      private List<MessageParam> dueSystemMessages() {
          List<MessageParam> due = new ArrayList<>();
          if (exitPending) {
              exitPending = false;
              modeAnnounced = false;
              due.add(systemMessage(MODE_EXIT));
          }
          if (modeOn) {
              if (!modeAnnounced) {
                  modeAnnounced = true;
                  turnsSinceReminder = 0;
                  due.add(systemMessage(MODE_ENTER));
              } else if (turnsSinceReminder >= TURNS_BETWEEN_REFRESHERS) {
                  turnsSinceReminder = 0;
                  due.add(systemMessage(MODE_REFRESH));
              }
          }
          return due;
      }

      private MessageParam systemMessage(String content) {
          return MessageParam.builder()
                  .role(MessageParam.Role.SYSTEM)
                  .content(content)
                  .build();
      }

      // Envía un turno de usuario por el bucle, ejecutando llamadas a herramientas hasta que el modelo se detenga.
      String turn(String userInput) throws InterruptedException {
          // Los mensajes de sistema en mitad de la conversación siguen al turno de usuario al que aplican, lo que mantiene
          // intacto el prefijo en caché que los precede.
          messages.add(MessageParam.builder().role(MessageParam.Role.USER).content(userInput).build());
          messages.addAll(dueSystemMessages());
          turnsSinceReminder++;

          for (int turn = 0; turn < MAX_MAIN_TURNS; turn++) {
              MessageCreateParams params = MessageCreateParams.builder()
                      .model(model)
                      .maxTokens(64000L)
                      .system(SYSTEM_PROMPT) // static for the whole session
                      .outputConfig(OutputConfig.builder().effort(EFFORT).build())
                      .addTool(WORKFLOW_TOOL)
                      .addTool(BASH_TOOL)
                      .messages(messages)
                      .build();
              MessageAccumulator accumulator = MessageAccumulator.create();
              try (var stream = client.messages().createStreaming(params, REQUEST_OPTIONS)) {
                  stream.stream().forEach(accumulator::accumulate);
              }
              Message response = accumulator.message();
              messages.add(response.toParam());

              StopReason stopReason = response.stopReason().orElse(null);
              if (StopReason.PAUSE_TURN.equals(stopReason)) {
                  continue;
              }
              if (!StopReason.TOOL_USE.equals(stopReason)) {
                  String text = response.content().stream()
                          .flatMap(block -> block.text().stream())
                          .map(TextBlock::text)
                          .collect(Collectors.joining());
                  if (StopReason.MAX_TOKENS.equals(stopReason)) {
                      // Descarta el mensaje truncado del asistente para que los turnos posteriores no se basen en él.
                      messages.removeLast();
                      text += "\n\n(warning: response was truncated at max_tokens)";
                  }
                  return text;
              }

              List<ToolUseBlock> toolUses = response.content().stream()
                      .flatMap(block -> block.toolUse().stream())
                      .toList();
              List<ContentBlockParam> toolResults = new ArrayList<>();
              for (ToolUseBlock toolUse : toolUses) {
                  ToolOutput result = switch (toolUse.name()) {
                      case "Workflow" -> runWorkflow(model,
                              toolInput(toolUse).getOrDefault("subtasks", JsonValue.from(List.of())));
                      case "bash" -> handleBashBlock(toolUse);
                      default -> new ToolOutput("unknown tool: " + toolUse.name(), true);
                  };
                  toolResults.add(ContentBlockParam.ofToolResult(ToolResultBlockParam.builder()
                          .toolUseId(toolUse.id())
                          .content(result.output())
                          .isError(result.isError())
                          .build()));
              }
              messages.add(MessageParam.builder()
                      .role(MessageParam.Role.USER)
                      .contentOfBlockParams(toolResults)
                      .build());
          }
          return "(hit the main loop turn limit before finishing)";
      }
  }
  ```

  ```php PHP
  /** An agent loop whose orchestration mode is toggled with mid-conversation system messages. */
  class ModeAgent
  {
      private array $messages = [];
      private bool $modeAnnounced = false;
      private bool $exitPending = false;
      private int $turnsSinceReminder = 0;

      public function __construct(
          private readonly Client $client,
          private readonly string $model,
          private bool $modeOn = true,
      ) {
      }

      /** Turn the mode on or off. The notice is delivered with the next user turn. */
      public function setMode(bool $modeOn): void
      {
          if ($modeOn === $this->modeOn) {
              return;
          }
          if ($modeOn) {
              $this->exitPending = false;
          } elseif ($this->modeAnnounced) {
              $this->exitPending = true;
          }
          $this->modeOn = $modeOn;
      }

      public function turn(string $userInput): string
      {
          // Los mensajes del sistema a mitad de conversación siguen al turno del usuario al que aplican, lo que mantiene
          // intacto el prefijo en caché que los precede.
          $this->messages[] = ['role' => 'user', 'content' => $userInput];
          array_push($this->messages, ...$this->dueSystemMessages());
          $this->turnsSinceReminder++;

          for ($turn = 0; $turn < MAX_MAIN_TURNS; $turn++) {
              $stream = $this->client->messages->createStream(
                  model: $this->model,
                  maxTokens: 64000,
                  system: SYSTEM_PROMPT, // static for the whole session
                  outputConfig: ['effort' => EFFORT],
                  tools: [WORKFLOW_TOOL, BASH_TOOL],
                  messages: $this->messages,
                  requestOptions: ['timeout' => REQUEST_TIMEOUT_SECONDS],
              );
              [$content, $stopReason] = drainMessageStream($stream);
              $this->messages[] = ['role' => 'assistant', 'content' => $content];

              if ($stopReason === 'pause_turn') {
                  continue;
              }
              if ($stopReason !== 'tool_use') {
                  $text = '';
                  foreach ($content as $block) {
                      if ($block instanceof \Anthropic\Messages\TextBlock) {
                          $text .= $block->text;
                      }
                  }
                  if ($stopReason === 'max_tokens') {
                      // Descarta el mensaje truncado del asistente para que los turnos posteriores no se basen en él.
                      array_pop($this->messages);
                      $text .= "\n\n(warning: response was truncated at max_tokens)";
                  }
                  return $text;
              }

              $toolResults = [];
              foreach ($content as $block) {
                  if (!$block instanceof \Anthropic\Messages\ToolUseBlock) {
                      continue;
                  }
                  if ($block->name === 'Workflow') {
                      [$output, $isError] =
                          runWorkflow($this->client, $this->model, $block->input['subtasks'] ?? []);
                  } elseif ($block->name === 'bash') {
                      [$output, $isError] = handleBashBlock($block);
                  } else {
                      $output = "unknown tool: {$block->name}";
                      $isError = true;
                  }
                  $toolResults[] = [
                      'type' => 'tool_result',
                      'tool_use_id' => $block->id,
                      'content' => $output,
                      'is_error' => $isError,
                  ];
              }
              $this->messages[] = ['role' => 'user', 'content' => $toolResults];
          }
          return '(hit the main loop turn limit before finishing)';
      }

      /**
       * System messages owed on this turn: an exit notice, the full mode text on entry,
       * or a one-line refresher every TURNS_BETWEEN_REFRESHERS user turns.
       */
      private function dueSystemMessages(): array
      {
          $due = [];
          if ($this->exitPending) {
              $this->exitPending = false;
              $this->modeAnnounced = false;
              $due[] = ['role' => 'system', 'content' => MODE_EXIT];
          }
          if ($this->modeOn) {
              if (!$this->modeAnnounced) {
                  $this->modeAnnounced = true;
                  $this->turnsSinceReminder = 0;
                  $due[] = ['role' => 'system', 'content' => MODE_ENTER];
              } elseif ($this->turnsSinceReminder >= TURNS_BETWEEN_REFRESHERS) {
                  $this->turnsSinceReminder = 0;
                  $due[] = ['role' => 'system', 'content' => MODE_REFRESH];
              }
          }
          return $due;
      }
  }
  ```

  ```ruby Ruby
  # Un bucle de agente cuyo modo de orquestación se alterna con mensajes del sistema a mitad de conversación.
  class ModeAgent
    def initialize(model, mode_on: true)
      @model = model
      @mode_on = mode_on
      @messages = []
      @mode_announced = false
      @exit_pending = false
      @turns_since_reminder = 0
    end

    # Activa o desactiva el modo. El aviso se entrega con el siguiente turno del usuario.
    def set_mode(mode_on)
      return if mode_on == @mode_on

      if mode_on
        @exit_pending = false
      else
        @exit_pending = true if @mode_announced
      end
      @mode_on = mode_on
    end

    def turn(user_input)
      # Los mensajes del sistema a mitad de conversación siguen al turno de usuario al que aplican, lo que mantiene
      # intacto el prefijo en caché que los precede.
      @messages << {role: "user", content: user_input}
      @messages.concat(due_system_messages)
      @turns_since_reminder += 1

      MAX_MAIN_TURNS.times do
        stream = CLIENT.messages.stream(
          model: @model,
          max_tokens: 64_000,
          system_: SYSTEM_PROMPT, # static for the whole session
          output_config: {effort: EFFORT},
          tools: [WORKFLOW_TOOL, BASH_TOOL],
          messages: @messages,
          request_options: {timeout: REQUEST_TIMEOUT_SECONDS}
        )
        response = stream.accumulated_message
        @messages << {role: "assistant", content: assistant_content_param(response.content)}

        next if response.stop_reason == :pause_turn

        unless response.stop_reason == :tool_use
          text = response.content.select { |block| block.type == :text }.map(&:text).join
          if response.stop_reason == :max_tokens
            # Descarta el mensaje truncado del asistente para que los turnos posteriores no se basen en él.
            @messages.pop
            text += "\n\n(warning: response was truncated at max_tokens)"
          end
          return text
        end

        tool_results = []
        response.content.each do |block|
          next unless block.type == :tool_use

          input = parse_tool_input(block.input)
          case block.name
          when "Workflow"
            output, is_error = run_workflow(@model, input["subtasks"] || [])
          when "bash"
            output, is_error = handle_bash_block(block)
          else
            output, is_error = "unknown tool: #{block.name}", true
          end
          tool_results << {
            type: "tool_result",
            tool_use_id: block.id,
            content: output,
            is_error: is_error
          }
        end
        @messages << {role: "user", content: tool_results}
      end
      "(hit the main loop turn limit before finishing)"
    end

    private

    # Mensajes del sistema pendientes en este turno: un aviso de salida, el texto completo del modo al entrar,
    # o un recordatorio de una línea cada TURNS_BETWEEN_REFRESHERS turnos de usuario.
    def due_system_messages
      due = []
      if @exit_pending
        @exit_pending = false
        @mode_announced = false
        due << {role: "system", content: MODE_EXIT}
      end
      if @mode_on
        if !@mode_announced
          @mode_announced = true
          @turns_since_reminder = 0
          due << {role: "system", content: MODE_ENTER}
        elsif @turns_since_reminder >= TURNS_BETWEEN_REFRESHERS
          @turns_since_reminder = 0
          due << {role: "system", content: MODE_REFRESH}
        end
      end
      due
    end
  end
  ```
</CodeGroup>

## Ejecútalo

<Warning>
  La herramienta bash de este ejemplo ejecuta comandos escritos por el modelo directamente en tu máquina, sin ningún entorno aislado, y la distribución en paralelo ejecuta varios de esos agentes a la vez. Ejecútalo en un directorio y un entorno que no te importe exponer, y agrega aislamiento antes de adaptarlo para cualquier uso que vaya más allá de la experimentación local.
</Warning>

<CodeGroup exclude="shell:cURL, shell:CLI">
  ```python Python
  if __name__ == "__main__":
      task = (
          sys.argv[1]
          if len(sys.argv) > 1
          else "Explore the current directory, then give a thorough review: what it does, "
          "code-quality issues, and concrete improvements."
      )
      agent = ModeAgent(MODEL)
      print(agent.turn(task))
      agent.set_mode(False)
      print(agent.turn("Briefly summarize what you found above, no fan-out needed."))
  ```

  ```typescript TypeScript
  const task =
    process.argv[2] ??
    "Explore the current directory, then give a thorough review: what it does, " +
      "code-quality issues, and concrete improvements.";
  const agent = new ModeAgent(MODEL);
  console.log(await agent.turn(task));
  agent.setMode(false);
  console.log(await agent.turn("Briefly summarize what you found above, no fan-out needed."));
  ```

  ```csharp C#
  var task = args.Length > 0
      ? args[0]
      : "Explore the current directory, then give a thorough review: what it does, "
          + "code-quality issues, and concrete improvements.";
  Console.WriteLine(await Turn(task));
  SetMode(false);
  Console.WriteLine(await Turn("Briefly summarize what you found above, no fan-out needed."));
  ```

  ```go Go
  func main() {
  	if err := run(context.Background()); err != nil {
  		log.Fatal(err)
  	}
  }

  func run(ctx context.Context) error {
  	if docTestMode {
  		defer os.RemoveAll(workDir)
  	}
  	task := "Explore the current directory, then give a thorough review: what it does, " +
  		"code-quality issues, and concrete improvements."
  	if len(os.Args) > 1 {
  		task = os.Args[1]
  	}
  	agent := newModeAgent(modelID)
  	answer, err := agent.turn(ctx, task)
  	if err != nil {
  		return err
  	}
  	fmt.Println(answer)

  	agent.setMode(false)
  	summary, err := agent.turn(ctx, "Briefly summarize what you found above, no fan-out needed.")
  	if err != nil {
  		return err
  	}
  	fmt.Println(summary)
  	return nil
  }
  ```

  ```java Java
  void main(String[] args) throws InterruptedException {
      String task = args.length > 0 ? args[0] : """
              Explore the current directory, then give a thorough review: what it does, \
              code-quality issues, and concrete improvements.""";
      ModeAgent agent = new ModeAgent(MODEL);
      IO.println(agent.turn(task));
      agent.setMode(false);
      IO.println(agent.turn("Briefly summarize what you found above, no fan-out needed."));
  }
  ```

  ```php PHP
  $task = $argv[1] ??
      'Explore the current directory, then give a thorough review: what it does, '
      . 'code-quality issues, and concrete improvements.';
  $agent = new ModeAgent($client, MODEL);
  echo $agent->turn($task), PHP_EOL;
  $agent->setMode(false);
  echo $agent->turn('Briefly summarize what you found above, no fan-out needed.'), PHP_EOL;
  ```

  ```ruby Ruby
  task = ARGV[0] ||
    "Explore the current directory, then give a thorough review: what it does, " \
    "code-quality issues, and concrete improvements."
  agent = ModeAgent.new(MODEL)
  puts agent.turn(task)
  agent.set_mode(false)
  puts agent.turn("Briefly summarize what you found above, no fan-out needed.")
  ```
</CodeGroup>

Inicia el ejemplo desde el directorio en el que quieres que trabajen los agentes, por ejemplo, la raíz de un repositorio que quieras revisar:

```bash
python orchestration_mode.py "Review this repository for flaky tests and propose fixes."
```

Con el modo activado, el modelo debería explorar con algunos comandos de bash, despachar la herramienta Workflow sin que se lo pidas y sintetizar los informes de los subagentes en una respuesta final. Las solicitudes triviales o conversacionales se resuelven sin subagentes, tal como indica el recordatorio.

## Hacia un arnés de producción

Este ejemplo es deliberadamente pequeño. Un arnés pensado para cargas de trabajo reales normalmente agregaría:

* **Scripts de orquestación en sandbox:** permite que el modelo emita un programa de orquestación breve (ramificaciones, bucles y pasos de reducción) y ejecútalo dentro de un intérprete aislado, en lugar de aceptar solo una lista plana de cadenas de subtareas.
* **Diario duradero:** reemplaza el archivo JSON local con un almacén que sobreviva a los reinicios del proceso y sea seguro con escritores concurrentes en varias máquinas.
* **Aplicación de presupuesto:** rastrea el total de subagentes lanzados en toda la sesión, no solo por llamada a Workflow, y niégate a superar un límite estricto para que un plan descontrolado no pueda agotar tu cuota.

Los patrones de este ejemplo (los recordatorios de modo, el consentimiento permanente en la descripción de la herramienta, el diario y una ola de verificación) se trasladan sin cambios; solo el sustrato de ejecución que los rodea se vuelve más robusto.

## Relacionado

<CardGroup cols={2}>
  <Card title="Mensajes del sistema a mitad de conversación" icon="message" href="https://platform.claude.com/docs/es/build-with-claude/mid-conversation-system-messages">
    El mecanismo que usan los recordatorios de modo y cómo interactúa con el almacenamiento en caché de prompts.
  </Card>

  <Card title="Esfuerzo" icon="gauge" href="https://platform.claude.com/docs/es/build-with-claude/effort">
    Los niveles de esfuerzo que acepta la API y cómo elegir uno.
  </Card>

  <Card title="Uso de herramientas con Claude" icon="wrench" href="https://platform.claude.com/docs/es/agents-and-tools/tool-use/overview">
    Definición de herramientas, manejo de llamadas a herramientas y resultados de herramientas.
  </Card>

  <Card title="Herramienta bash" icon="terminal" href="https://platform.claude.com/docs/es/agents-and-tools/tool-use/bash-tool">
    La herramienta bash definida por Anthropic que este ejemplo ejecuta localmente.
  </Card>
</CardGroup>