{"meta":{"title":"Crea tu primera aplicación con tecnología Copilot","intro":"En este tutorial, usará el SDK de Copilot para compilar un asistente de línea de comandos. Empezarás con los conceptos básicos, añadirás respuestas en tiempo real y, a continuación, añadirás herramientas personalizadas, lo que le dará a Copilot la capacidad de invocar tu código.","product":"GitHub Copilot","breadcrumbs":[{"href":"/es/copilot","title":"GitHub Copilot"},{"href":"/es/copilot/how-tos","title":"Procedimientos"},{"href":"/es/copilot/how-tos/copilot-sdk","title":"SDK de Copilot"},{"href":"/es/copilot/how-tos/copilot-sdk/getting-started","title":"Introducción"}],"documentType":"article"},"body":"# Crea tu primera aplicación con tecnología Copilot\n\nEn este tutorial, usará el SDK de Copilot para compilar un asistente de línea de comandos. Empezarás con los conceptos básicos, añadirás respuestas en tiempo real y, a continuación, añadirás herramientas personalizadas, lo que le dará a Copilot la capacidad de invocar tu código.\n\n\n\n\n\n**Lo que vas a construir:**\n\n```text\nYou: What's the weather like in Seattle?\nCopilot: Let me check the weather for Seattle...\n Currently 62°F and cloudy with a chance of rain.\n Typical Seattle weather!\n\nYou: How about Tokyo?\nCopilot: In Tokyo it's 75°F and sunny. Great day to be outside!\n```\n\n## Prerequisites\n\nAntes de comenzar, asegúrese de que tiene:\n\n* **GitHub Copilot CLI** instalada y autenticada (los SDK de Node.js, Python y .NET proporcionan automáticamente la CLI; consulte [Configuración predeterminada (CLI agrupada)](/es/copilot/how-tos/copilot-sdk/setup/bundled-cli). Necesario para Go, Java y Rust, a menos que usen sus funciones de empaquetado de la CLI a nivel de aplicación).\n* Su entorno de ejecución de lenguaje preferido:\n * **Node.js** 20+ o **Python** 3.11+ o **Go** 1.24+ o **Rust** 1.94+ o **Java** 17+ o **.NET** 8.0+\n\nCompruebe que la CLI funciona:\n\n```bash\ncopilot --version\n```\n\n## Paso 1: instalar el SDK\n\n
\n
TypeScript
\n\nEn primer lugar, cree un nuevo directorio e inicialice el proyecto:\n\n```bash\nmkdir copilot-demo && cd copilot-demo\nnpm init -y --init-type module\n```\n\nA continuación, instale el SDK y el ejecutor de TypeScript:\n\n```bash\nnpm install @github/copilot-sdk tsx\n```\n\n
\n\n
Python
\n\n```bash\npip install github-copilot-sdk\n```\n\n
\n\n
Go
\n\nEn primer lugar, cree un nuevo directorio e inicialice el módulo:\n\n```bash\nmkdir copilot-demo && cd copilot-demo\ngo mod init copilot-demo\n```\n\nA continuación, instale el SDK:\n\n```bash\ngo get github.com/github/copilot-sdk/go\n```\n\n
\n\n
Rust
\n\nEn primer lugar, cree una nueva caja binaria:\n\n```bash\ncargo new copilot-demo && cd copilot-demo\n```\n\nA continuación, instale el SDK y las dependencias directas usadas por los ejemplos:\n\n```bash\ncargo add github-copilot-sdk --features derive\n# Used by #[tokio::main] and tokio::spawn\ncargo add tokio --features rt-multi-thread,macros\n# Used by custom-tool parameter derives later in this guide\ncargo add serde --features derive\ncargo add schemars\n```\n\n
\n\n
.NET
\n\nEn primer lugar, cree un proyecto de consola:\n\n```bash\ndotnet new console -n CopilotDemo && cd CopilotDemo\n```\n\nA continuación, agregue el SDK:\n\n```bash\ndotnet add package GitHub.Copilot.SDK\n```\n\n
\n\n
Java
\n\nEn primer lugar, cree un nuevo directorio e inicialice el proyecto.\n\n**Maven**: añade a tu `pom.xml`\n\n```xml\n\n com.github\n copilot-sdk-java\n ${copilot.sdk.version}\n\n```\n\n**Gradle**—añadir a tu `build.gradle`:\n\n```groovy\nimplementation 'com.github:copilot-sdk-java:${copilotSdkVersion}'\n```\n\n
\n\n
\n\n## Paso 2: enviar el primer mensaje\n\nCree un nuevo archivo y agregue el código siguiente. Esta es la manera más sencilla de usar el SDK, aproximadamente 5 líneas de código.\n\n
\n
TypeScript
\n\nCreación de `index.ts`:\n\n```typescript\nimport { CopilotClient } from \"@github/copilot-sdk\";\n\nconst client = new CopilotClient();\nconst session = await client.createSession({ model: \"auto\" });\n\nconst response = await session.sendAndWait({ prompt: \"What is 2 + 2?\" });\nconsole.log(response?.data.content);\n\nawait client.stop();\nprocess.exit(0);\n```\n\nEjecútelo:\n\n```bash\nnpx tsx index.ts\n```\n\n
\n\n
Python
\n\nCreación de `main.py`:\n\n```python\nimport asyncio\nfrom copilot import CopilotClient\nfrom copilot.session import PermissionHandler\n\nasync def main():\n client = CopilotClient()\n await client.start()\n\n session = await client.create_session(on_permission_request=PermissionHandler.approve_all, model=\"auto\")\n response = await session.send_and_wait(\"What is 2 + 2?\")\n print(response.data.content)\n\n await client.stop()\n\nasyncio.run(main())\n```\n\nEjecútelo:\n\n```bash\npython main.py\n```\n\n
\n\n
Go
\n\nCreación de `main.go`:\n\n```golang\npackage main\n\nimport (\n \"context\"\n \"fmt\"\n \"log\"\n \"os\"\n\n copilot \"github.com/github/copilot-sdk/go\"\n)\n\nfunc main() {\n ctx := context.Background()\n client := copilot.NewClient(nil)\n if err := client.Start(ctx); err != nil {\n log.Fatal(err)\n }\n defer client.Stop()\n\n session, err := client.CreateSession(ctx, &copilot.SessionConfig{Model: \"auto\"})\n if err != nil {\n log.Fatal(err)\n }\n\n response, err := session.SendAndWait(ctx, copilot.MessageOptions{Prompt: \"What is 2 + 2?\"})\n if err != nil {\n log.Fatal(err)\n }\n\n if d, ok := response.Data.(*copilot.AssistantMessageData); ok {\n fmt.Println(d.Content)\n }\n os.Exit(0)\n}\n```\n\nEjecútelo:\n\n```bash\ngo run main.go\n```\n\n
\n\n
Rust
\n\nCreación de `src/main.rs`:\n\n```rust\nuse std::sync::Arc;\nuse std::time::Duration;\n\nuse github_copilot_sdk::handler::ApproveAllHandler;\nuse github_copilot_sdk::{Client, ClientOptions, MessageOptions, SessionConfig};\n\n#[tokio::main]\nasync fn main() -> Result<(), Box> {\n let client = Client::start(ClientOptions::default()).await?;\n let session = client\n .create_session(SessionConfig::default().with_permission_handler(Arc::new(ApproveAllHandler)))\n .await?;\n\n let response = session\n .send_and_wait(\n MessageOptions::new(\"What is 2 + 2?\").with_wait_timeout(Duration::from_secs(120)),\n )\n .await?;\n\n if let Some(event) = response {\n if let Some(content) = event.data.get(\"content\").and_then(|value| value.as_str()) {\n println!(\"{content}\");\n }\n }\n\n session.disconnect().await?;\n client.stop().await?;\n Ok(())\n}\n```\n\nEjecútelo:\n\n```bash\ncargo run\n```\n\n
\n\n
.NET
\n\nCree un nuevo proyecto de consola y agréguelo a `Program.cs`:\n\n```csharp\nusing GitHub.Copilot;\n\nawait using var client = new CopilotClient();\nawait using var session = await client.CreateSessionAsync(new SessionConfig\n{\n Model = \"auto\",\n OnPermissionRequest = PermissionHandler.ApproveAll\n});\n\nvar response = await session.SendAndWaitAsync(new MessageOptions { Prompt = \"What is 2 + 2?\" });\nConsole.WriteLine(response?.Data.Content);\n```\n\nEjecútelo:\n\n```bash\ndotnet run\n```\n\n
\n\n
Java
\n\nCreación de `HelloCopilot.java`:\n\n\n\n```java\nimport com.github.copilot.CopilotClient;\nimport com.github.copilot.rpc.*;\n\npublic class HelloCopilot {\n public static void main(String[] args) throws Exception {\n try (var client = new CopilotClient()) {\n client.start().get();\n\n var session = client.createSession(\n new SessionConfig()\n .setModel(\"auto\")\n .setOnPermissionRequest(PermissionHandler.APPROVE_ALL)\n ).get();\n\n var response = session.sendAndWait(\n new MessageOptions().setPrompt(\"What is 2 + 2?\")\n ).get();\n\n System.out.println(response.getData().content());\n\n client.stop().get();\n }\n }\n}\n```\n\nEjecútelo:\n\n```bash\njavac -cp copilot-sdk.jar HelloCopilot.java && java -cp .:copilot-sdk.jar HelloCopilot\n```\n\n
\n\n
\n\n**Debería ver:**\n\n```text\n4\n```\n\n¡Felicidades! Acaba de crear su primera aplicación con tecnología Copilot.\n\n## Paso 3: agregar respuestas de streaming\n\nAhora mismo, tienes que esperar a que se muestre la respuesta completa antes de ver nada. Vamos a hacer que sea interactivo mediante el streaming de la respuesta a medida que se genera.\n\n
\n
TypeScript
\n\nActualizar `index.ts`:\n\n```typescript\nimport { CopilotClient } from \"@github/copilot-sdk\";\n\nconst client = new CopilotClient();\nconst session = await client.createSession({\n model: \"auto\",\n streaming: true,\n});\n\n// Listen for response chunks\nsession.on(\"assistant.message_delta\", (event) => {\n process.stdout.write(event.data.deltaContent);\n});\nsession.on(\"session.idle\", () => {\n console.log(); // New line when done\n});\n\nawait session.sendAndWait({ prompt: \"Tell me a short joke\" });\n\nawait client.stop();\nprocess.exit(0);\n```\n\n
\n\n
Python
\n\nActualizar `main.py`:\n\n```python\nimport asyncio\nimport sys\nfrom copilot import CopilotClient\nfrom copilot.session import PermissionHandler\nfrom copilot.session_events import SessionEventType\n\nasync def main():\n client = CopilotClient()\n await client.start()\n\n session = await client.create_session(on_permission_request=PermissionHandler.approve_all, model=\"auto\", streaming=True)\n\n # Listen for response chunks\n def handle_event(event):\n if event.type == SessionEventType.ASSISTANT_MESSAGE_DELTA:\n sys.stdout.write(event.data.delta_content)\n sys.stdout.flush()\n if event.type == SessionEventType.SESSION_IDLE:\n print() # New line when done\n\n session.on(handle_event)\n\n await session.send_and_wait(\"Tell me a short joke\")\n\n await client.stop()\n\nasyncio.run(main())\n```\n\n
\n\n
Go
\n\nActualizar `main.go`:\n\n```golang\npackage main\n\nimport (\n \"context\"\n \"fmt\"\n \"log\"\n \"os\"\n\n copilot \"github.com/github/copilot-sdk/go\"\n)\n\nfunc main() {\n ctx := context.Background()\n client := copilot.NewClient(nil)\n if err := client.Start(ctx); err != nil {\n log.Fatal(err)\n }\n defer client.Stop()\n\n session, err := client.CreateSession(ctx, &copilot.SessionConfig{\n Model: \"auto\",\n Streaming: copilot.Bool(true),\n })\n if err != nil {\n log.Fatal(err)\n }\n\n // Listen for response chunks\n session.On(func(event copilot.SessionEvent) {\n switch d := event.Data.(type) {\n case *copilot.AssistantMessageDeltaData:\n fmt.Print(d.DeltaContent)\n case *copilot.SessionIdleData:\n _ = d\n fmt.Println()\n }\n })\n\n _, err = session.SendAndWait(ctx, copilot.MessageOptions{Prompt: \"Tell me a short joke\"})\n if err != nil {\n log.Fatal(err)\n }\n os.Exit(0)\n}\n```\n\n
\n\n
Rust
\n\nActualizar `src/main.rs`:\n\n```rust\nuse std::io::{self, Write};\nuse std::sync::Arc;\nuse std::time::Duration;\n\nuse github_copilot_sdk::handler::ApproveAllHandler;\nuse github_copilot_sdk::{Client, ClientOptions, MessageOptions, SessionConfig};\n\n#[tokio::main]\nasync fn main() -> Result<(), Box> {\n let client = Client::start(ClientOptions::default()).await?;\n\n let mut config = SessionConfig::default();\n config.streaming = Some(true);\n let session = client\n .create_session(config.with_permission_handler(Arc::new(ApproveAllHandler)))\n .await?;\n\n // Listen for response chunks\n let mut events = session.subscribe();\n tokio::spawn(async move {\n while let Ok(event) = events.recv().await {\n match event.event_type.as_str() {\n \"assistant.message_delta\" => {\n if let Some(text) =\n event.data.get(\"deltaContent\").and_then(|value| value.as_str())\n {\n print!(\"{text}\");\n io::stdout().flush().ok();\n }\n }\n \"assistant.message\" => println!(),\n _ => {}\n }\n }\n });\n\n session\n .send_and_wait(\n MessageOptions::new(\"Tell me a short joke\")\n .with_wait_timeout(Duration::from_secs(120)),\n )\n .await?;\n\n session.disconnect().await?;\n client.stop().await?;\n Ok(())\n}\n```\n\n
\n\n
.NET
\n\nActualizar `Program.cs`:\n\n```csharp\nusing GitHub.Copilot;\n\nawait using var client = new CopilotClient();\nawait using var session = await client.CreateSessionAsync(new SessionConfig\n{\n Model = \"auto\",\n OnPermissionRequest = PermissionHandler.ApproveAll,\n Streaming = true,\n});\n\n// Listen for response chunks\nsession.On(ev =>\n{\n if (ev is AssistantMessageDeltaEvent deltaEvent)\n {\n Console.Write(deltaEvent.Data.DeltaContent);\n }\n if (ev is SessionIdleEvent)\n {\n Console.WriteLine();\n }\n});\n\nawait session.SendAndWaitAsync(new MessageOptions { Prompt = \"Tell me a short joke\" });\n```\n\n
\n\n
Java
\n\nActualizar `HelloCopilot.java`:\n\n\n\n```java\nimport com.github.copilot.CopilotClient;\nimport com.github.copilot.rpc.*;\n\npublic class HelloCopilot {\n public static void main(String[] args) throws Exception {\n try (var client = new CopilotClient()) {\n client.start().get();\n\n var session = client.createSession(\n new SessionConfig()\n .setModel(\"auto\")\n .setStreaming(true)\n .setOnPermissionRequest(PermissionHandler.APPROVE_ALL)\n ).get();\n\n // Listen for response chunks\n session.on(AssistantMessageDeltaEvent.class, delta -> {\n System.out.print(delta.getData().deltaContent());\n });\n session.on(SessionIdleEvent.class, idle -> {\n System.out.println(); // New line when done\n });\n\n session.sendAndWait(\n new MessageOptions().setPrompt(\"Tell me a short joke\")\n ).get();\n\n client.stop().get();\n }\n }\n}\n```\n\n
\n\n
\n\nVuelva a ejecutar el código. Verá que la respuesta aparece palabra por palabra.\n\n### Métodos de suscripción de eventos\n\nEl SDK proporciona métodos para suscribirse a eventos de sesión:\n\n| Método | Description |\n| ------------------------ | ----------------------------------------------------------------------------------------------------------------------- |\n| `on(handler)` | Suscribirse a todos los eventos; devuelve la función de cancelación de suscripción |\n| `on(eventType, handler)` | Suscribirse a un tipo específico de evento (solo Node.js/TypeScript); devuelve una función para cancelar la suscripción |\n| `subscribe()` | Suscribirse a todos los eventos (Rust); filtrar por `event_type` |\n\n
\n
TypeScript
\n\n```typescript\n// Subscribe to all events\nconst unsubscribeAll = session.on((event) => {\n console.log(\"Event:\", event.type);\n});\n\n// Subscribe to specific event type\nconst unsubscribeIdle = session.on(\"session.idle\", (event) => {\n console.log(\"Session is idle\");\n});\n\n// Later, to unsubscribe:\nunsubscribeAll();\nunsubscribeIdle();\n```\n\n
\n\n
Python
\n\n```python\n# Subscribe to all events\nunsubscribe = session.on(lambda event: print(f\"Event: {event.type}\"))\n\n# Filter by event type in your handler\ndef handle_event(event):\n if event.type == SessionEventType.SESSION_IDLE:\n print(\"Session is idle\")\n elif event.type == SessionEventType.ASSISTANT_MESSAGE:\n print(f\"Message: {event.data.content}\")\n\nunsubscribe = session.on(handle_event)\n\n# Later, to unsubscribe:\nunsubscribe()\n```\n\n
\n\n
Go
\n\n```golang\n// Subscribe to all events\nunsubscribe := session.On(func(event copilot.SessionEvent) {\n fmt.Println(\"Event:\", event.Type)\n})\n\n// Filter by event type in your handler\nsession.On(func(event copilot.SessionEvent) {\n switch d := event.Data.(type) {\n case *copilot.SessionIdleData:\n _ = d\n fmt.Println(\"Session is idle\")\n case *copilot.AssistantMessageData:\n fmt.Println(\"Message:\", d.Content)\n }\n})\n\n// Later, to unsubscribe:\nunsubscribe()\n```\n\n
\n\n
Rust
\n\n```rust\nlet mut events = session.subscribe();\n\ntokio::spawn(async move {\n while let Ok(event) = events.recv().await {\n println!(\"Event: {}\", event.event_type);\n\n match event.event_type.as_str() {\n \"session.idle\" => println!(\"Session is idle\"),\n \"assistant.message\" => {\n if let Some(content) = event.data.get(\"content\").and_then(|value| value.as_str()) {\n println!(\"Message: {content}\");\n }\n }\n _ => {}\n }\n }\n});\n```\n\n
\n\n
.NET
\n\n```csharp\n// Subscribe to all events\nvar unsubscribe = session.On(ev => Console.WriteLine($\"Event: {ev.Type}\"));\n\n// Filter by event type using pattern matching\nsession.On(ev =>\n{\n switch (ev)\n {\n case SessionIdleEvent:\n Console.WriteLine(\"Session is idle\");\n break;\n case AssistantMessageEvent msg:\n Console.WriteLine($\"Message: {msg.Data.Content}\");\n break;\n }\n});\n\n// Later, to unsubscribe:\nunsubscribe.Dispose();\n```\n\n
\n\n
Java
\n\n\n\n```java\n// Subscribe to all events\nvar unsubscribe = session.on(event -> {\n System.out.println(\"Event: \" + event.getType());\n});\n\n// Subscribe to a specific event type\nsession.on(AssistantMessageEvent.class, msg -> {\n System.out.println(\"Message: \" + msg.getData().content());\n});\n\nsession.on(SessionIdleEvent.class, idle -> {\n System.out.println(\"Session is idle\");\n});\n\n// Later, to unsubscribe:\nunsubscribe.close();\n```\n\n
\n\n
\n\n## Paso 4: agregar una herramienta personalizada\n\nY ahora viene la parte potente. Vamos a darle a Copilot la capacidad de llamar a tu código definiendo una herramienta personalizada. Crearemos una herramienta de búsqueda meteorológica sencilla.\n\n
\n
TypeScript
\n\nActualizar `index.ts`:\n\n```typescript\nimport { CopilotClient, defineTool } from \"@github/copilot-sdk\";\n\n// Define a tool that Copilot can call\nconst getWeather = defineTool(\"get_weather\", {\n description: \"Get the current weather for a city\",\n parameters: {\n type: \"object\",\n properties: {\n city: { type: \"string\", description: \"The city name\" },\n },\n required: [\"city\"],\n },\n handler: async (args: { city: string }) => {\n const { city } = args;\n // In a real app, you'd call a weather API here\n const conditions = [\"sunny\", \"cloudy\", \"rainy\", \"partly cloudy\"];\n const temp = Math.floor(Math.random() * 30) + 50;\n const condition = conditions[Math.floor(Math.random() * conditions.length)];\n return { city, temperature: `${temp}°F`, condition };\n },\n});\n\nconst client = new CopilotClient();\nconst session = await client.createSession({\n model: \"auto\",\n streaming: true,\n tools: [getWeather],\n});\n\nsession.on(\"assistant.message_delta\", (event) => {\n process.stdout.write(event.data.deltaContent);\n});\n\nsession.on(\"session.idle\", () => {\n console.log(); // New line when done\n});\n\nawait session.sendAndWait({\n prompt: \"What's the weather like in Seattle and Tokyo?\",\n});\n\nawait client.stop();\nprocess.exit(0);\n```\n\n
\n\n
Python
\n\nActualizar `main.py`:\n\n```python\nimport asyncio\nimport random\nimport sys\nfrom copilot import CopilotClient\nfrom copilot.session import PermissionHandler\nfrom copilot.tools import define_tool\nfrom copilot.session_events import SessionEventType\nfrom pydantic import BaseModel, Field\n\n# Define the parameters for the tool using Pydantic\nclass GetWeatherParams(BaseModel):\n city: str = Field(description=\"The name of the city to get weather for\")\n\n# Define a tool that Copilot can call\n@define_tool(description=\"Get the current weather for a city\")\nasync def get_weather(params: GetWeatherParams) -> dict:\n city = params.city\n # In a real app, you'd call a weather API here\n conditions = [\"sunny\", \"cloudy\", \"rainy\", \"partly cloudy\"]\n temp = random.randint(50, 80)\n condition = random.choice(conditions)\n return {\"city\": city, \"temperature\": f\"{temp}°F\", \"condition\": condition}\n\nasync def main():\n client = CopilotClient()\n await client.start()\n\n session = await client.create_session(on_permission_request=PermissionHandler.approve_all, model=\"auto\", streaming=True, tools=[get_weather])\n\n def handle_event(event):\n if event.type == SessionEventType.ASSISTANT_MESSAGE_DELTA:\n sys.stdout.write(event.data.delta_content)\n sys.stdout.flush()\n if event.type == SessionEventType.SESSION_IDLE:\n print()\n\n session.on(handle_event)\n\n await session.send_and_wait(\"What's the weather like in Seattle and Tokyo?\")\n\n await client.stop()\n\nasyncio.run(main())\n```\n\n
\n\n
Go
\n\nActualizar `main.go`:\n\n```golang\npackage main\n\nimport (\n \"context\"\n \"fmt\"\n \"log\"\n \"math/rand\"\n \"os\"\n\n copilot \"github.com/github/copilot-sdk/go\"\n)\n\n// Define the parameter type\ntype WeatherParams struct {\n City string `json:\"city\" jsonschema:\"The city name\"`\n}\n\n// Define the return type\ntype WeatherResult struct {\n City string `json:\"city\"`\n Temperature string `json:\"temperature\"`\n Condition string `json:\"condition\"`\n}\n\nfunc main() {\n ctx := context.Background()\n\n // Define a tool that Copilot can call\n getWeather := copilot.DefineTool(\n \"get_weather\",\n \"Get the current weather for a city\",\n func(params WeatherParams, inv copilot.ToolInvocation) (WeatherResult, error) {\n // In a real app, you'd call a weather API here\n conditions := []string{\"sunny\", \"cloudy\", \"rainy\", \"partly cloudy\"}\n temp := rand.Intn(30) + 50\n condition := conditions[rand.Intn(len(conditions))]\n return WeatherResult{\n City: params.City,\n Temperature: fmt.Sprintf(\"%d°F\", temp),\n Condition: condition,\n }, nil\n },\n )\n\n client := copilot.NewClient(nil)\n if err := client.Start(ctx); err != nil {\n log.Fatal(err)\n }\n defer client.Stop()\n\n session, err := client.CreateSession(ctx, &copilot.SessionConfig{\n Model: \"auto\",\n Streaming: copilot.Bool(true),\n Tools: []copilot.Tool{getWeather},\n })\n if err != nil {\n log.Fatal(err)\n }\n\n session.On(func(event copilot.SessionEvent) {\n switch d := event.Data.(type) {\n case *copilot.AssistantMessageDeltaData:\n fmt.Print(d.DeltaContent)\n case *copilot.SessionIdleData:\n _ = d\n fmt.Println()\n }\n })\n\n _, err = session.SendAndWait(ctx, copilot.MessageOptions{\n Prompt: \"What's the weather like in Seattle and Tokyo?\",\n })\n if err != nil {\n log.Fatal(err)\n }\n os.Exit(0)\n}\n```\n\n
\n\n
Rust
\n\nActualizar `src/main.rs`:\n\n```rust\nuse std::io::{self, Write};\nuse std::sync::Arc;\nuse std::time::Duration;\n\nuse github_copilot_sdk::handler::ApproveAllHandler;\nuse github_copilot_sdk::tool::{define_tool, JsonSchema};\nuse github_copilot_sdk::{Client, ClientOptions, MessageOptions, SessionConfig, ToolResult};\nuse serde::Deserialize;\n\n#[derive(Deserialize, JsonSchema)]\nstruct GetWeatherParams {\n city: String,\n}\n\n#[tokio::main]\nasync fn main() -> Result<(), Box> {\n // Define a tool that Copilot can call\n let tools = vec![define_tool(\n \"get_weather\",\n \"Get the current weather for a city\",\n |_inv, params: GetWeatherParams| async move {\n Ok(ToolResult::Text(format!(\n \"{}: 62°F and sunny\",\n params.city\n )))\n },\n )];\n\n let client = Client::start(ClientOptions::default()).await?;\n\n let mut config = SessionConfig::default();\n config.streaming = Some(true);\n let session = client\n .create_session(\n config\n .with_tools(tools)\n .with_permission_handler(Arc::new(ApproveAllHandler)),\n )\n .await?;\n\n let mut events = session.subscribe();\n tokio::spawn(async move {\n while let Ok(event) = events.recv().await {\n match event.event_type.as_str() {\n \"assistant.message_delta\" => {\n if let Some(text) =\n event.data.get(\"deltaContent\").and_then(|value| value.as_str())\n {\n print!(\"{text}\");\n io::stdout().flush().ok();\n }\n }\n \"assistant.message\" => println!(),\n _ => {}\n }\n }\n });\n\n session\n .send_and_wait(\n MessageOptions::new(\"What's the weather like in Seattle and Tokyo?\")\n .with_wait_timeout(Duration::from_secs(120)),\n )\n .await?;\n\n session.disconnect().await?;\n client.stop().await?;\n Ok(())\n}\n```\n\n
\n\n
.NET
\n\nActualizar `Program.cs`:\n\n```csharp\nusing GitHub.Copilot;\nusing Microsoft.Extensions.AI;\nusing System.ComponentModel;\n\nawait using var client = new CopilotClient();\n\n// Define a tool that Copilot can call\nvar getWeather = CopilotTool.DefineTool(\n ([Description(\"The city name\")] string city) =>\n {\n // In a real app, you'd call a weather API here\n var conditions = new[] { \"sunny\", \"cloudy\", \"rainy\", \"partly cloudy\" };\n var temp = Random.Shared.Next(50, 80);\n var condition = conditions[Random.Shared.Next(conditions.Length)];\n return new { city, temperature = $\"{temp}°F\", condition };\n },\n factoryOptions: new AIFunctionFactoryOptions\n {\n Name = \"get_weather\",\n Description = \"Get the current weather for a city\",\n }\n);\n\nawait using var session = await client.CreateSessionAsync(new SessionConfig\n{\n Model = \"auto\",\n OnPermissionRequest = PermissionHandler.ApproveAll,\n Streaming = true,\n Tools = [getWeather],\n});\n\nsession.On(ev =>\n{\n if (ev is AssistantMessageDeltaEvent deltaEvent)\n {\n Console.Write(deltaEvent.Data.DeltaContent);\n }\n if (ev is SessionIdleEvent)\n {\n Console.WriteLine();\n }\n});\n\nawait session.SendAndWaitAsync(new MessageOptions\n{\n Prompt = \"What's the weather like in Seattle and Tokyo?\",\n});\n```\n\n
\n\n
Java
\n\nActualizar `HelloCopilot.java`:\n\n\n\n```java\nimport com.github.copilot.CopilotClient;\nimport com.github.copilot.rpc.*;\n\nimport java.util.List;\nimport java.util.Map;\nimport java.util.Random;\nimport java.util.concurrent.CompletableFuture;\n\npublic class HelloCopilot {\n public static void main(String[] args) throws Exception {\n var random = new Random();\n var conditions = List.of(\"sunny\", \"cloudy\", \"rainy\", \"partly cloudy\");\n\n // Define a tool that Copilot can call\n var getWeather = ToolDefinition.create(\n \"get_weather\",\n \"Get the current weather for a city\",\n Map.of(\n \"type\", \"object\",\n \"properties\", Map.of(\n \"city\", Map.of(\"type\", \"string\", \"description\", \"The city name\")\n ),\n \"required\", List.of(\"city\")\n ),\n invocation -> {\n var city = (String) invocation.getArguments().get(\"city\");\n var temp = random.nextInt(30) + 50;\n var condition = conditions.get(random.nextInt(conditions.size()));\n return CompletableFuture.completedFuture(Map.of(\n \"city\", city,\n \"temperature\", temp + \"°F\",\n \"condition\", condition\n ));\n }\n );\n\n try (var client = new CopilotClient()) {\n client.start().get();\n\n var session = client.createSession(\n new SessionConfig()\n .setModel(\"auto\")\n .setStreaming(true)\n .setTools(List.of(getWeather))\n .setOnPermissionRequest(PermissionHandler.APPROVE_ALL)\n ).get();\n\n session.on(AssistantMessageDeltaEvent.class, delta -> {\n System.out.print(delta.getData().deltaContent());\n });\n session.on(SessionIdleEvent.class, idle -> {\n System.out.println();\n });\n\n session.sendAndWait(\n new MessageOptions().setPrompt(\"What's the weather like in Seattle and Tokyo?\")\n ).get();\n\n client.stop().get();\n }\n }\n}\n```\n\n
\n\n
\n\nEjecútelo y verá que Copilot usa la herramienta para obtener datos del tiempo y, a continuación, responde con los resultados.\n\n## Paso 5: crear un asistente interactivo\n\nVamos a ponerlo todo juntos en un asistente interactivo útil:\n\n
\n
TypeScript
\n\n```typescript\nimport { CopilotClient, defineTool } from \"@github/copilot-sdk\";\nimport * as readline from \"readline\";\n\nconst getWeather = defineTool(\"get_weather\", {\n description: \"Get the current weather for a city\",\n parameters: {\n type: \"object\",\n properties: {\n city: { type: \"string\", description: \"The city name\" },\n },\n required: [\"city\"],\n },\n handler: async ({ city }) => {\n const conditions = [\"sunny\", \"cloudy\", \"rainy\", \"partly cloudy\"];\n const temp = Math.floor(Math.random() * 30) + 50;\n const condition = conditions[Math.floor(Math.random() * conditions.length)];\n return { city, temperature: `${temp}°F`, condition };\n },\n});\n\nconst client = new CopilotClient();\nconst session = await client.createSession({\n model: \"auto\",\n streaming: true,\n tools: [getWeather],\n});\n\nsession.on(\"assistant.message_delta\", (event) => {\n process.stdout.write(event.data.deltaContent);\n});\n\nconst rl = readline.createInterface({\n input: process.stdin,\n output: process.stdout,\n});\n\nconsole.log(\"🌤️ Weather Assistant (type 'exit' to quit)\");\nconsole.log(\" Try: 'What's the weather in Paris?'\\n\");\n\nconst prompt = () => {\n rl.question(\"You: \", async (input) => {\n if (input.toLowerCase() === \"exit\") {\n await client.stop();\n rl.close();\n return;\n }\n\n process.stdout.write(\"Assistant: \");\n await session.sendAndWait({ prompt: input });\n console.log(\"\\n\");\n prompt();\n });\n};\n\nprompt();\n```\n\nEjecute con:\n\n```bash\nnpx tsx weather-assistant.ts\n```\n\n
\n\n
Python
\n\nCreación de `weather_assistant.py`:\n\n```python\nimport asyncio\nimport random\nimport sys\nfrom copilot import CopilotClient\nfrom copilot.session import PermissionHandler\nfrom copilot.tools import define_tool\nfrom copilot.session_events import SessionEventType\nfrom pydantic import BaseModel, Field\n\nclass GetWeatherParams(BaseModel):\n city: str = Field(description=\"The name of the city to get weather for\")\n\n@define_tool(description=\"Get the current weather for a city\")\nasync def get_weather(params: GetWeatherParams) -> dict:\n city = params.city\n conditions = [\"sunny\", \"cloudy\", \"rainy\", \"partly cloudy\"]\n temp = random.randint(50, 80)\n condition = random.choice(conditions)\n return {\"city\": city, \"temperature\": f\"{temp}°F\", \"condition\": condition}\n\nasync def main():\n client = CopilotClient()\n await client.start()\n\n session = await client.create_session(on_permission_request=PermissionHandler.approve_all, model=\"auto\", streaming=True, tools=[get_weather])\n\n def handle_event(event):\n if event.type == SessionEventType.ASSISTANT_MESSAGE_DELTA:\n sys.stdout.write(event.data.delta_content)\n sys.stdout.flush()\n\n session.on(handle_event)\n\n print(\"🌤️ Weather Assistant (type 'exit' to quit)\")\n print(\" Try: 'What's the weather in Paris?' or 'Compare weather in NYC and LA'\\n\")\n\n while True:\n try:\n user_input = input(\"You: \")\n except EOFError:\n break\n\n if user_input.lower() == \"exit\":\n break\n\n sys.stdout.write(\"Assistant: \")\n await session.send_and_wait(user_input)\n print(\"\\n\")\n\n await client.stop()\n\nasyncio.run(main())\n```\n\nEjecute con:\n\n```bash\npython weather_assistant.py\n```\n\n
\n\n
Go
\n\nCreación de `weather-assistant.go`:\n\n```golang\npackage main\n\nimport (\n \"bufio\"\n \"context\"\n \"fmt\"\n \"log\"\n \"math/rand\"\n \"os\"\n \"strings\"\n\n copilot \"github.com/github/copilot-sdk/go\"\n)\n\ntype WeatherParams struct {\n City string `json:\"city\" jsonschema:\"The city name\"`\n}\n\ntype WeatherResult struct {\n City string `json:\"city\"`\n Temperature string `json:\"temperature\"`\n Condition string `json:\"condition\"`\n}\n\nfunc main() {\n ctx := context.Background()\n\n getWeather := copilot.DefineTool(\n \"get_weather\",\n \"Get the current weather for a city\",\n func(params WeatherParams, inv copilot.ToolInvocation) (WeatherResult, error) {\n conditions := []string{\"sunny\", \"cloudy\", \"rainy\", \"partly cloudy\"}\n temp := rand.Intn(30) + 50\n condition := conditions[rand.Intn(len(conditions))]\n return WeatherResult{\n City: params.City,\n Temperature: fmt.Sprintf(\"%d°F\", temp),\n Condition: condition,\n }, nil\n },\n )\n\n client := copilot.NewClient(nil)\n if err := client.Start(ctx); err != nil {\n log.Fatal(err)\n }\n defer client.Stop()\n\n session, err := client.CreateSession(ctx, &copilot.SessionConfig{\n Model: \"auto\",\n Streaming: copilot.Bool(true),\n Tools: []copilot.Tool{getWeather},\n })\n if err != nil {\n log.Fatal(err)\n }\n\n session.On(func(event copilot.SessionEvent) {\n switch d := event.Data.(type) {\n case *copilot.AssistantMessageDeltaData:\n fmt.Print(d.DeltaContent)\n case *copilot.SessionIdleData:\n _ = d\n fmt.Println()\n }\n })\n\n fmt.Println(\"🌤️ Weather Assistant (type 'exit' to quit)\")\n fmt.Println(\" Try: 'What's the weather in Paris?' or 'Compare weather in NYC and LA'\\n\")\n\n scanner := bufio.NewScanner(os.Stdin)\n for {\n fmt.Print(\"You: \")\n if !scanner.Scan() {\n break\n }\n input := scanner.Text()\n if strings.ToLower(input) == \"exit\" {\n break\n }\n\n fmt.Print(\"Assistant: \")\n _, err = session.SendAndWait(ctx, copilot.MessageOptions{Prompt: input})\n if err != nil {\n fmt.Fprintf(os.Stderr, \"Error: %v\\n\", err)\n break\n }\n fmt.Println()\n }\n if err := scanner.Err(); err != nil {\n fmt.Fprintf(os.Stderr, \"Input error: %v\\n\", err)\n }\n}\n```\n\nEjecute con:\n\n```bash\ngo run weather-assistant.go\n```\n\n
\n\n
Rust
\n\nCreación de `src/main.rs`:\n\n```rust\nuse std::io::{self, BufRead, Write};\nuse std::sync::Arc;\nuse std::time::Duration;\n\nuse github_copilot_sdk::handler::ApproveAllHandler;\nuse github_copilot_sdk::tool::{define_tool, JsonSchema};\nuse github_copilot_sdk::{Client, ClientOptions, MessageOptions, SessionConfig, ToolResult};\nuse serde::Deserialize;\n\n#[derive(Deserialize, JsonSchema)]\nstruct GetWeatherParams {\n city: String,\n}\n\nfn read_line() -> Option {\n let stdin = io::stdin();\n let mut line = String::new();\n stdin.lock().read_line(&mut line).ok()?;\n if line.is_empty() {\n return None;\n }\n Some(line.trim_end_matches(&['\\n', '\\r'][..]).to_string())\n}\n\n#[tokio::main]\nasync fn main() -> Result<(), Box> {\n let tools = vec![define_tool(\n \"get_weather\",\n \"Get the current weather for a city\",\n |_inv, params: GetWeatherParams| async move {\n Ok(ToolResult::Text(format!(\n \"{}: 62°F and sunny\",\n params.city\n )))\n },\n )];\n\n let client = Client::start(ClientOptions::default()).await?;\n\n let mut config = SessionConfig::default();\n config.streaming = Some(true);\n let session = client\n .create_session(\n config\n .with_tools(tools)\n .with_permission_handler(Arc::new(ApproveAllHandler)),\n )\n .await?;\n\n let mut events = session.subscribe();\n tokio::spawn(async move {\n while let Ok(event) = events.recv().await {\n match event.event_type.as_str() {\n \"assistant.message_delta\" => {\n if let Some(text) =\n event.data.get(\"deltaContent\").and_then(|value| value.as_str())\n {\n print!(\"{text}\");\n io::stdout().flush().ok();\n }\n }\n \"assistant.message\" => println!(),\n _ => {}\n }\n }\n });\n\n println!(\"Weather Assistant (type 'exit' to quit)\");\n println!(\"Try: 'What's the weather in Paris?' or 'Compare weather in NYC and LA'\\n\");\n\n loop {\n print!(\"You: \");\n io::stdout().flush().ok();\n\n let Some(input) = read_line() else { break };\n if input.eq_ignore_ascii_case(\"exit\") {\n break;\n }\n\n print!(\"Assistant: \");\n io::stdout().flush().ok();\n session\n .send_and_wait(MessageOptions::new(input).with_wait_timeout(Duration::from_secs(120)))\n .await?;\n println!();\n }\n\n session.disconnect().await?;\n client.stop().await?;\n Ok(())\n}\n```\n\nEjecute con:\n\n```bash\ncargo run\n```\n\n
\n\n
.NET
\n\nCree un nuevo proyecto de consola y actualice `Program.cs`:\n\n```csharp\nusing GitHub.Copilot;\nusing Microsoft.Extensions.AI;\nusing System.ComponentModel;\n\n// Define the weather tool\nvar getWeather = CopilotTool.DefineTool(\n ([Description(\"The city name\")] string city) =>\n {\n var conditions = new[] { \"sunny\", \"cloudy\", \"rainy\", \"partly cloudy\" };\n var temp = Random.Shared.Next(50, 80);\n var condition = conditions[Random.Shared.Next(conditions.Length)];\n return new { city, temperature = $\"{temp}°F\", condition };\n },\n factoryOptions: new AIFunctionFactoryOptions\n {\n Name = \"get_weather\",\n Description = \"Get the current weather for a city\",\n });\n\nawait using var client = new CopilotClient();\nawait using var session = await client.CreateSessionAsync(new SessionConfig\n{\n Model = \"auto\",\n OnPermissionRequest = PermissionHandler.ApproveAll,\n Streaming = true,\n Tools = [getWeather]\n});\n\n// Listen for response chunks\nsession.On(ev =>\n{\n if (ev is AssistantMessageDeltaEvent deltaEvent)\n {\n Console.Write(deltaEvent.Data.DeltaContent);\n }\n if (ev is SessionIdleEvent)\n {\n Console.WriteLine();\n }\n});\n\nConsole.WriteLine(\"🌤️ Weather Assistant (type 'exit' to quit)\");\nConsole.WriteLine(\" Try: 'What's the weather in Paris?' or 'Compare weather in NYC and LA'\\n\");\n\nwhile (true)\n{\n Console.Write(\"You: \");\n var input = Console.ReadLine();\n\n if (string.IsNullOrEmpty(input) || input.Equals(\"exit\", StringComparison.OrdinalIgnoreCase))\n {\n break;\n }\n\n Console.Write(\"Assistant: \");\n await session.SendAndWaitAsync(new MessageOptions { Prompt = input });\n Console.WriteLine(\"\\n\");\n}\n```\n\nEjecute con:\n\n```bash\ndotnet run\n```\n\n
\n\n
Java
\n\nCreación de `WeatherAssistant.java`:\n\n\n\n```java\nimport com.github.copilot.CopilotClient;\nimport com.github.copilot.rpc.*;\n\nimport java.util.List;\nimport java.util.Map;\nimport java.util.Random;\nimport java.util.Scanner;\nimport java.util.concurrent.CompletableFuture;\n\npublic class WeatherAssistant {\n public static void main(String[] args) throws Exception {\n var random = new Random();\n var conditions = List.of(\"sunny\", \"cloudy\", \"rainy\", \"partly cloudy\");\n\n var getWeather = ToolDefinition.create(\n \"get_weather\",\n \"Get the current weather for a city\",\n Map.of(\n \"type\", \"object\",\n \"properties\", Map.of(\n \"city\", Map.of(\"type\", \"string\", \"description\", \"The city name\")\n ),\n \"required\", List.of(\"city\")\n ),\n invocation -> {\n var city = (String) invocation.getArguments().get(\"city\");\n var temp = random.nextInt(30) + 50;\n var condition = conditions.get(random.nextInt(conditions.size()));\n return CompletableFuture.completedFuture(Map.of(\n \"city\", city,\n \"temperature\", temp + \"°F\",\n \"condition\", condition\n ));\n }\n );\n\n try (var client = new CopilotClient()) {\n client.start().get();\n\n var session = client.createSession(\n new SessionConfig()\n .setModel(\"auto\")\n .setStreaming(true)\n .setOnPermissionRequest(request ->\n CompletableFuture.completedFuture(PermissionDecision.allow())\n )\n .setTools(List.of(getWeather))\n ).get();\n\n session.on(AssistantMessageDeltaEvent.class, delta -> {\n System.out.print(delta.getData().deltaContent());\n });\n session.on(SessionIdleEvent.class, idle -> {\n System.out.println();\n });\n\n System.out.println(\"🌤️ Weather Assistant (type 'exit' to quit)\");\n System.out.println(\" Try: 'What's the weather in Paris?' or 'Compare weather in NYC and LA'\\n\");\n\n var scanner = new Scanner(System.in);\n while (true) {\n System.out.print(\"You: \");\n if (!scanner.hasNextLine()) break;\n var input = scanner.nextLine();\n if (input.equalsIgnoreCase(\"exit\")) break;\n\n System.out.print(\"Assistant: \");\n session.sendAndWait(\n new MessageOptions().setPrompt(input)\n ).get();\n System.out.println(\"\\n\");\n }\n\n client.stop().get();\n }\n }\n}\n```\n\nEjecute con:\n\n```bash\njavac -cp copilot-sdk.jar WeatherAssistant.java && java -cp .:copilot-sdk.jar WeatherAssistant\n```\n\n
\n\n
\n\n**Sesión de ejemplo:**\n\n```text\n🌤️ Weather Assistant (type 'exit' to quit)\n Try: 'What's the weather in Paris?' or 'Compare weather in NYC and LA'\n\nYou: What's the weather in Seattle?\nAssistant: Let me check the weather for Seattle...\nIt's currently 62°F and cloudy in Seattle.\n\nYou: How about Tokyo and London?\nAssistant: I'll check both cities for you:\n- Tokyo: 75°F and sunny\n- London: 58°F and rainy\n\nYou: exit\n```\n\n¡Has creado un asistente con una herramienta personalizada que Copilot puede invocar!\n\n## Funcionamiento de las herramientas\n\nCuando defines una herramienta, le estás diciendo a Copilot:\n\n1. **Qué hace la herramienta** (descripción)\n2. **Qué parámetros necesita** (esquema)\n3. **Qué código se va a ejecutar** (controlador)\n\nCopilot decide cuándo llamar a la herramienta en función de la pregunta del usuario. Cuando lo haga:\n\n1. Copilot envía una solicitud de llamada de herramienta con los parámetros\n2. El SDK ejecuta la función del controlador.\n3. El resultado se devuelve a Copilot\n4. Copilot incorpora el resultado en su respuesta\n\n## ¿Qué sigue?\n\nAhora que tiene los conceptos básicos, estas son características más eficaces para explorar:\n\n### Conexión a servidores MCP\n\nLos servidores MCP (Protocolo de contexto de modelo) proporcionan herramientas pregeneradas. Conéctese al servidor MCP de GitHub para conceder a Copilot acceso a repositorios, problemas y solicitudes de incorporación de cambios:\n\n```typescript\nconst session = await client.createSession({\n mcpServers: {\n github: {\n type: \"http\",\n url: \"https://api.githubcopilot.com/mcp/\",\n },\n },\n});\n```\n\n📖\n\\*\\*\n[Uso de servidores MCP con el SDK de GitHub Copilot](/es/copilot/how-tos/copilot-sdk/features/mcp)\\*\\* : obtenga información sobre servidores locales frente a remotos, todas las opciones de configuración y la solución de problemas.\n\n### Creación de agentes personalizados\n\nDefina roles de IA especializados para tareas específicas:\n\n```typescript\nconst session = await client.createSession({\n customAgents: [{\n name: \"pr-reviewer\",\n displayName: \"PR Reviewer\",\n description: \"Reviews pull requests for best practices\",\n prompt: \"You are an expert code reviewer. Focus on security, performance, and maintainability.\",\n }],\n});\n```\n\n> \\[!TIP]\n> También puede establecer `agent: \"pr-reviewer\"` en la configuración de sesión para seleccionar previamente este agente desde el principio. Consulte [autotitle](/es/copilot/how-tos/copilot-sdk/features/custom-agents#selecting-an-agent-at-session-creation) para obtener más información.\n\n### Personalización del mensaje del sistema\n\nControle el comportamiento y la personalidad de la inteligencia artificial anexando instrucciones:\n\n```typescript\nconst session = await client.createSession({\n systemMessage: {\n content: \"You are a helpful assistant for our engineering team. Always be concise.\",\n },\n});\n```\n\nPara un control más detallado, use `mode: \"customize\"` para anular secciones individuales del prompt del sistema sin dejar de preservar el resto:\n\n```typescript\nconst session = await client.createSession({\n systemMessage: {\n mode: \"customize\",\n sections: {\n tone: { action: \"replace\", content: \"Respond in a warm, professional tone. Be thorough in explanations.\" },\n code_change_rules: { action: \"remove\" },\n guidelines: { action: \"append\", content: \"\\n* Always cite data sources\" },\n },\n content: \"Focus on financial analysis and reporting.\",\n },\n});\n```\n\nIdentificadores de sección disponibles: `preamble`, `identity`, `tone`, `tool_efficiency`, `code_change_rules`, `guidelines`, `environment_context`, `safety`, `tool_instructions`, `custom_instructions`, `last_instructions`, `runtime_instructions`.\n\n`identity` y `tool_instructions` son *grupos* de secciones: tienen como destino una colección de subsecciones relacionadas como una unidad. Utilice `preamble` para apuntar solo al preámbulo de la identidad sin afectar a sus subsecciones hermanas.\n\nCada anulación admite cinco acciones: `replace`, `remove`, `append`, `prepend`, y `preserve`. La acción `preserve` es una operación nula que excluye una sección direccionable individualmente de un `remove` en un grupo (por ejemplo, mantener `tone` al eliminar el grupo `identity`). Los identificadores de sección desconocidos se gestionan correctamente: el contenido de las anulaciones de `replace`/`append`/`prepend` se añade a las instrucciones adicionales, y las anulaciones de `remove` se omiten silenciosamente.\n\nConsulte los READMEs del SDK específicos del lenguaje para obtener ejemplos en [TypeScript](https://github.com/github/copilot-sdk/tree/main/nodejs/README.md), [Python](https://github.com/github/copilot-sdk/tree/main/python/README.md), [Go](https://github.com/github/copilot-sdk/tree/main/go/README.md), [Rust](https://github.com/github/copilot-sdk/tree/main/rust/README.md), [Java](https://github.com/github/copilot-sdk/tree/main/java/README.md) y [C#](https://github.com/github/copilot-sdk/tree/main/dotnet/README.md).\n\n## Conexión a un servidor de la CLI externo\n\nDe forma predeterminada, el SDK administra automáticamente el ciclo de vida del proceso de la CLI de Copilot, iniciando y deteniendo la CLI según sea necesario. Sin embargo, también puede ejecutar la CLI en modo de servidor por separado y hacer que el SDK se conecte a él. Esto puede ser útil para:\n\n* **Depuración**: Mantenga la interfaz de línea de comandos en ejecución entre reinicios del SDK para inspeccionar los registros\n* **Uso compartido de recursos**: varios clientes del SDK pueden conectarse al mismo servidor de la CLI\n* **Desarrollo**: ejecute la CLI con la configuración personalizada o en un entorno diferente.\n\n### Ejecución de la CLI en modo de servidor\n\nInicie la CLI en modo de servidor con la `--headless` marca y, opcionalmente, especifique un puerto:\n\n```bash\ncopilot --headless --port 4321\n```\n\nSi no especifica un puerto, la CLI elegirá un puerto disponible aleatoriamente.\n\nDe forma predeterminada, el servidor headless solo acepta conexiones de loopback (`127.0.0.1`), por lo que el SDK debe ejecutarse en la misma máquina. Para aceptar conexiones desde otros hosts (por ejemplo, al ejecutar la CLI en un contenedor o en un servidor independiente), vincúlelo a una dirección distinta de la de loopback con `--host`:\n\n```bash\n# Listen on all interfaces\ncopilot --headless --host 0.0.0.0 --port 4321\n```\n\n> \\[!WARNING]\n> Exponer el servidor sin encabezado en una dirección que no sea de bucle invertido hace que sea accesible por cualquier persona que pueda enrutar a esa dirección. Complételo con controles de red (cortafuegos, una red privada, proxy inverso) y mecanismos de autenticación adecuados para su entorno.\n\n### Conexión del SDK al servidor externo\n\nUna vez que la CLI se ejecuta en modo de servidor, configure el cliente del SDK para conectarse a él mediante la opción \"cli url\":\n\n
\n
TypeScript
\n\n```typescript\nimport { CopilotClient, approveAll } from \"@github/copilot-sdk\";\n\nconst client = new CopilotClient({\n cliUrl: \"localhost:4321\"\n});\n\n// Use the client normally\nconst session = await client.createSession({ onPermissionRequest: approveAll });\n// ...\n```\n\n
\n\n
Python
\n\n```python\nfrom copilot import CopilotClient, RuntimeConnection\nfrom copilot.session import PermissionHandler\n\nclient = CopilotClient(connection=RuntimeConnection.for_uri(\"localhost:4321\"))\nawait client.start()\n\n# Use the client normally\nsession = await client.create_session(on_permission_request=PermissionHandler.approve_all)\n# ...\n```\n\n
\n\n
Go
\n\n```golang\nimport copilot \"github.com/github/copilot-sdk/go\"\n\nclient := copilot.NewClient(&copilot.ClientOptions{\n Connection: copilot.URIConnection{URL: \"localhost:4321\"},\n})\n\nif err := client.Start(ctx); err != nil {\n log.Fatal(err)\n}\ndefer client.Stop()\n\n// Use the client normally\nsession, err := client.CreateSession(ctx, &copilot.SessionConfig{\n OnPermissionRequest: copilot.PermissionHandler.ApproveAll,\n})\n// ...\n```\n\n
\n\n
Rust
\n\n```rust\nuse std::sync::Arc;\n\nuse github_copilot_sdk::handler::ApproveAllHandler;\nuse github_copilot_sdk::{Client, ClientOptions, SessionConfig, Transport};\n\nlet mut options = ClientOptions::default();\noptions.transport = Transport::External {\n host: \"localhost\".to_string(),\n port: 4321,\n connection_token: None,\n};\nlet client = Client::start(options).await?;\n\n// Use the client normally\nlet session = client\n .create_session(SessionConfig::default().with_permission_handler(Arc::new(ApproveAllHandler)))\n .await?;\n// ...\n```\n\n
\n\n
.NET
\n\n```csharp\nusing GitHub.Copilot;\n\nusing var client = new CopilotClient(new CopilotClientOptions\n{\n Connection = RuntimeConnection.ForUri(\"localhost:4321\"),\n});\n\n// Use the client normally\nawait using var session = await client.CreateSessionAsync(new()\n{\n OnPermissionRequest = PermissionHandler.ApproveAll\n});\n// ...\n```\n\n
\n\n
Java
\n\n```java\nimport com.github.copilot.CopilotClient;\nimport com.github.copilot.rpc.*;\n\nvar client = new CopilotClient(\n new CopilotClientOptions().setCliUrl(\"localhost:4321\")\n);\nclient.start().get();\n\n// Use the client normally\nvar session = client.createSession(\n new SessionConfig().setOnPermissionRequest(PermissionHandler.APPROVE_ALL)\n).get();\n// ...\n```\n\n
\n\n
\n\n**Nota:** Cuando se proporciona `cli_url` / `cliUrl` o el `URIConnection` de Go, o Rust usa `Transport::External`, el SDK no iniciará ni administrará un proceso de CLI; solo se conectará al servidor existente en la URL especificada.\n\n## Telemetría y observabilidad\n\nEl SDK de Copilot admite [OpenTelemetry](https://opentelemetry.io/) para el seguimiento distribuido. Proporcione una `telemetry` configuración al cliente para habilitar la exportación de seguimiento desde el proceso de la CLI y la propagación automática del [contexto de seguimiento de W3C](https://www.w3.org/TR/trace-context/) entre el SDK y la CLI.\n\n### Habilitación de telemetría\n\nPasa una configuración `telemetry` (o `Telemetry`) al crear el cliente. Esta es la opción de aceptación voluntaria; no se necesita ningún indicador de «habilitado» por separado.\n\n
\n
TypeScript
\n\n\n\n```typescript\nimport { CopilotClient } from \"@github/copilot-sdk\";\n\nconst client = new CopilotClient({\n telemetry: {\n otlpEndpoint: \"http://localhost:4318\",\n },\n});\n```\n\nDependencia de par opcional: `@opentelemetry/api`\n\n
\n\n
Python
\n\n\n\n```python\nfrom copilot import CopilotClient, CopilotClientOptions\n\nclient = CopilotClient(CopilotClientOptions(\n telemetry={\n \"otlp_endpoint\": \"http://localhost:4318\",\n },\n))\n```\n\nInstalación con extras de telemetría: `pip install copilot-sdk[telemetry]` (proporciona `opentelemetry-api`)\n\n
\n\n
Go
\n\n\n\n```golang\nclient := copilot.NewClient(&copilot.ClientOptions{\n Telemetry: &copilot.TelemetryConfig{\n OTLPEndpoint: \"http://localhost:4318\",\n },\n})\n```\n\nDependencia: `go.opentelemetry.io/otel`\n\n
\n\n
Rust
\n\n\n\n```rust\nuse github_copilot_sdk::{Client, ClientOptions, OtelExporterType, TelemetryConfig};\n\nlet mut options = ClientOptions::default();\noptions.telemetry = Some(\n TelemetryConfig::new()\n .with_exporter_type(OtelExporterType::OtlpHttp)\n .with_otlp_endpoint(\"http://localhost:4318\"),\n);\nlet client = Client::start(options).await?;\n```\n\nSin dependencias adicionales: el SDK inserta variables de entorno de telemetría para el proceso de la CLI generado.\n\n
\n\n
.NET
\n\n\n\n```csharp\nvar client = new CopilotClient(new CopilotClientOptions\n{\n Telemetry = new TelemetryConfig\n {\n OtlpEndpoint = \"http://localhost:4318\",\n },\n});\n```\n\nSin dependencias adicionales: usa `System.Diagnostics.Activity` integrado.\n\n
\n\n
Java
\n\n\n\n```java\nimport com.github.copilot.CopilotClient;\nimport com.github.copilot.rpc.*;\n\nvar client = new CopilotClient(new CopilotClientOptions()\n .setTelemetry(new TelemetryConfig()\n .setOtlpEndpoint(\"http://localhost:4318\")));\n```\n\nDependencia: `io.opentelemetry:opentelemetry-api`\n\n
\n\n
\n\n### Opciones de TelemetryConfig\n\n| Opción | Node.js | Python | Ir | Óxido | Java | .NET | Description |\n| ------------------------- | ---------------- | ----------------- | ---------------- | ----------------- | ---------------- | ---------------- | -------------------------------------------------------------------------------- |\n| Punto de conexión de OTLP | `otlpEndpoint` | `otlp_endpoint` | `OTLPEndpoint` | `otlp_endpoint` | `otlpEndpoint` | `OtlpEndpoint` | Dirección URL del punto de conexión HTTP de OTLP |\n| Protocolo OTLP | `otlpProtocol` | `otlp_protocol` | `OTLPProtocol` | `otlp_protocol` | `otlpProtocol` | `OtlpProtocol` | Protocolo HTTP de OTLP para todas las señales: `\"http/json\"` o `\"http/protobuf\"` |\n| Ruta del archivo | `filePath` | `file_path` | `FilePath` | `file_path` | `filePath` | `FilePath` | Ruta de acceso de archivo para salida de seguimiento de líneas JSON |\n| Tipo de exportador | `exporterType` | `exporter_type` | `ExporterType` | `exporter_type` | `exporterType` | `ExporterType` | |\n| `\"otlp-http\"` o `\"file\"` | | | | | | | |\n| Nombre de origen | `sourceName` | `source_name` | `SourceName` | `source_name` | `sourceName` | `SourceName` | Nombre del ámbito de instrumentación |\n| Captura de contenido | `captureContent` | `capture_content` | `CaptureContent` | `capture_content` | `captureContent` | `CaptureContent` | Si se va a capturar el contenido del mensaje |\n\nEl campo de protocolo OTLP configura el exportador `\"otlp-http\"` de la CLI para todas las señales. Déjelo sin definir para usar el valor predeterminado de la CLI, o configúrelo como `\"http/protobuf\"` para exportar en formato protobuf a través de HTTP.\n\n### Exportación de archivos\n\nPara escribir las trazas en un archivo local en lugar de en un punto de conexión OTLP:\n\n\n\n```typescript\nconst client = new CopilotClient({\n telemetry: {\n filePath: \"./traces.jsonl\",\n exporterType: \"file\",\n },\n});\n```\n\n### Propagación del contexto de rastreo\n\nEl contexto de seguimiento se propaga automáticamente; no se necesita instrumentación manual:\n\n* **SDK → CLI**: los encabezados `traceparent` y `tracestate` del span o la actividad actuales se incluyen en las llamadas RPC `session.create`, `session.resume` y `session.send`.\n* **CLI → SDK**: cuando la CLI invoca controladores de herramientas, el contexto de seguimiento del intervalo de la CLI se propaga para que el código de la herramienta se ejecute bajo el intervalo primario correcto.\n\n📖\n\\*\\*\n[Instrumentación de OpenTelemetry para el SDK de Copilot](/es/copilot/how-tos/copilot-sdk/observability/opentelemetry)\\*\\*: opciones de TelemetryConfig, propagación de contexto de seguimiento y dependencias por lenguaje.\n\n## Aprende más\n\n* [Autenticación](/es/copilot/how-tos/copilot-sdk/auth/authenticate): GitHub OAuth, variables de entorno y BYOK\n* [BYOK (traiga su propia clave)](/es/copilot/how-tos/copilot-sdk/auth/byok) - Usa tus propias claves de API de Microsoft Foundry, OpenAI, etc.\n* [ Referencia del SDK deNode.js](https://github.com/github/copilot-sdk/tree/main/nodejs/README.md)\n* Referencia del SDK de [Python](https://github.com/github/copilot-sdk/tree/main/python/README.md)\n* [Referencia del SDK de Go](https://github.com/github/copilot-sdk/tree/main/go/README.md)\n* [Referencia del SDK de Rust](https://github.com/github/copilot-sdk/tree/main/rust/README.md)\n* [Referencia del SDK de .NET](https://github.com/github/copilot-sdk/tree/main/dotnet/README.md)\n* Referencia del SDK de [Java](https://github.com/github/copilot-sdk/tree/main/java/README.md)\n* [Uso de servidores MCP con el SDK de GitHub Copilot](/es/copilot/how-tos/copilot-sdk/features/mcp) : integración de herramientas externas mediante el protocolo de contexto de modelo\n* [GitHub documentación del servidor MCP](https://github.com/github/github-mcp-server)\n* [Directorio de servidores MCP](https://github.com/modelcontextprotocol/servers) : exploración de más servidores MCP\n* [Instrumentación de OpenTelemetry para el SDK de Copilot](/es/copilot/how-tos/copilot-sdk/observability/opentelemetry) - TelemetryConfig, propagación del contexto de seguimiento y dependencias específicas de cada lenguaje\n\n**¡Lo hiciste!** Ha aprendido los conceptos básicos del SDK de GitHub Copilot:\n\n* ✅ Creación de un cliente y una sesión\n* ✅ Envío de mensajes y recepción de respuestas\n* ✅ Transmisión de resultados en tiempo real\n* ✅ Definir herramientas personalizadas que Copilot puede invocar\n\n¡Ahora vamos a construir algo increíble!\n🚀"}