AGENT SKILL

cloud-functions

Complete guide for CloudBase cloud functions development - supports both Event Functions (Node.js) and HTTP Functions (multi-language Web services). Covers runtime selection, deployment, logging, invocation, scf_bootstrap, SSE, WebSocket, and HTTP access configuration.

License 许可
MIT
最近更新
2026-11-10
Stars
★ 1
Forks
0

01 前言 Front Matter

name
cloud-functions
alwaysApply
false
description
Complete guide for CloudBase cloud functions development - supports both Event Functions (Node.js) and HTTP Functions (multi-language Web services). Covers runtime selection, deployment, logging, invocation, scf_bootstrap, SSE, WebSocket, and HTTP access configuration.

02 资源文件 1

03 SKILL.md 正文

Cloud Functions Development

Activation Contract

Use this first when

  • The task is to create, update, deploy, inspect, or debug a CloudBase Event Function or HTTP Function.

Read before writing code if

  • The request mentions runtime, HTTP function, scf_bootstrap, function logs, or function deployment.

Then also read

  • Auth setup or provider-related backend work -> ../auth-tool/SKILL.md
  • AI in functions -> ../ai-model-nodejs/SKILL.md

Do NOT use for

  • CloudRun container services or Web authentication UI implementation.

Common mistakes / gotchas

  • Picking the wrong function type and trying to compensate later.
  • Mixing Event Function code shape (exports.main(event, context)) with HTTP Function code shape (req / res on port 9000).
  • Treating HTTP Access as the implementation model for HTTP Functions. HTTP Access is a gateway configuration for Event Functions, not the HTTP Function runtime model.
  • Forgetting that runtime cannot be changed after creation.
  • Using cloud functions as the first solution for Web login.
  • Forgetting that HTTP Functions do NOT auto-install dependencies — always include node_modules.
  • Forgetting to add app.use(express.json()) or equivalent body parsing middleware for POST/PUT requests.
  • Leaving HTTP handlers without sending a response (causes request timeout).
  • Not returning proper HTTP status codes (400, 401, 405) for error cases.

Minimal checklist

Overview

Use this skill when developing, deploying, and managing CloudBase cloud functions. CloudBase supports two types of cloud functions:

  • Event Functions (普通云函数): Traditional serverless functions triggered by events (SDK calls, timers)
  • HTTP Functions (HTTP 云函数): Web service functions triggered by HTTP requests, supporting multiple languages

Writing mode at a glance

  • If the request is for a normal CloudBase function triggered by SDK calls, timers, or other events, write an Event Function with exports.main = async (event, context) => {}.
  • If the request is for an HTTP endpoint, REST API, SSE, or WebSocket service, write an HTTP Function that listens on port 9000 and handles req / res.
  • If the user mentions HTTP access for an existing Event Function, keep the Event Function code shape and add gateway access separately.

When to use this skill

Use this skill for cloud function operations when you need to:

  • Create and deploy Event Functions (Node.js)
  • Create and deploy HTTP Functions (Node.js/Python/Go/Java)
  • Understand runtime limitations and selection
  • Query function logs and monitor execution
  • Invoke cloud functions from applications
  • Configure HTTP access for cloud functions
  • Implement SSE (Server-Sent Events) or WebSocket

Do NOT use for:

  • CloudRun backend services (use cloudrun-development skill)
  • Complex container-based services (use cloudrun-development skill)
  • Database operations (use database skills)

How to use this skill (for a coding agent)

  1. Choose the right function type

    • Event Function: For SDK calls, scheduled tasks, event-driven scenarios
    • HTTP Function: For Web APIs, REST services, SSE/WebSocket, multi-language support
    • If the prompt says "HTTP 云函数", use the HTTP Function model first; if it says "触发器云函数" or omits HTTP, default to the Event Function model.
  2. Understand runtime limitations

    • Runtime CANNOT be changed after function creation
    • Must select correct runtime during initial creation
    • If runtime needs to change, must delete and recreate function
  3. Write code AND deploy — do not stop after writing files

    • CRITICAL: After writing function code, you MUST also deploy it using manageFunctions. Creating files locally without deploying is incomplete.
    • Use manageFunctions(action="createFunction") to create AND deploy in one step
    • Provide functionRootPath as the parent directory of the function folder (e.g., /project/cloudfunctions/ not /project/cloudfunctions/myFunction/)
    • Preferred MCP Tools: Use queryFunctions for reads, manageFunctions(action="createFunction") for creation, and manageFunctions(action="updateFunctionCode") for code updates
    • Legacy compatibility: If older prompts mention createFunction / updateFunctionCode, map them to the manageFunctions actions above
    • CLI: Use tcb fn deploy (Event) or tcb fn deploy --httpFn (HTTP) only as a fallback when MCP tools are unavailable
    • In agent / non-interactive runs, never default to CLI login flows for deployment; keep the flow on manageFunctions
    • For HTTP functions, create or update them through manageFunctions with func.type="HTTP" as the primary path
    • HTTP Functions require scf_bootstrap file in the function directory
    • Provide correct functionRootPath (parent directory of function folder)
  4. Query logs properly

    • Use queryFunctions(action="listFunctionLogs") for log list (basic info)
    • Use queryFunctions(action="getFunctionLogDetail") with RequestId for detailed logs
    • Legacy compatibility: Treat getFunctionLogs / getFunctionLogDetail as the two queryFunctions actions above
    • Note time range limitations (max 1 day interval)
  5. Handle legacy tool names safely

    • Prefer the converged entrances: queryFunctions, manageFunctions, queryGateway, manageGateway
    • If old names appear in historical docs or prompts, convert them before acting:
      • getFunctionList -> queryFunctions
      • createFunction -> manageFunctions(action="createFunction")
      • updateFunctionCode -> manageFunctions(action="updateFunctionCode")
      • updateFunctionConfig -> manageFunctions(action="updateFunctionConfig")
      • getFunctionLogs -> queryFunctions(action="listFunctionLogs")
      • getFunctionLogDetail -> queryFunctions(action="getFunctionLogDetail")
      • manageFunctionTriggers -> manageFunctions(action="createFunctionTrigger"|"deleteFunctionTrigger")
      • readFunctionLayers -> queryFunctions(action="listLayers"|"listLayerVersions"|"getLayerVersionDetail"|"listFunctionLayers")
      • writeFunctionLayers -> manageFunctions(action="createLayerVersion"|"deleteLayerVersion"|"attachLayer"|"detachLayer"|"updateFunctionLayers")
      • createFunctionHTTPAccess -> manageGateway(action="createAccess")

Function Types Comparison

Feature Event Function (普通云函数) HTTP Function (HTTP 云函数)
Trigger Event-driven (SDK, timer) HTTP request
Entry Point exports.main = async (event, context) => {} Web Server (Express/Flask/Gin etc.)
Port No port required Must listen on port 9000
Bootstrap Not required Requires scf_bootstrap
Connection Short connection Supports long connection
Languages Node.js only Node.js, Python, Go, Java
Protocols N/A HTTP, SSE, WebSocket
Use Cases Event processing, scheduled tasks Web APIs, REST services, real-time streaming

Core Knowledge - Event Functions

Runtime Environment

⚠️ CRITICAL: Runtime cannot be modified after function creation

Once a cloud function is created with a specific runtime, the runtime cannot be changed. If you need a different runtime:

  1. Delete the existing function
  2. Create a new function with the desired runtime

Supported Node.js Runtimes:

  • Nodejs18.15 (Default, Recommended)
  • Nodejs16.13
  • Nodejs14.18
  • Nodejs12.16
  • Nodejs10.15
  • Nodejs8.9

Runtime Selection Guidelines:

  • Use Nodejs18.15 for new projects (default, most modern)
  • Choose older versions only if dependencies require specific Node.js versions
  • Consider security updates and support lifecycle
  • Test thoroughly with selected runtime before deployment

Event Function Structure

Event functions require:

  1. Function Directory: Contains function code

    • Must have index.js (or specified entry file)
    • Must export handler: exports.main = async (event, context) => {}
    • Include package.json with dependencies
  2. Function Root Path: Parent directory containing function directories

    • Example: If function is at /project/cloudfunctions/myFunction/
    • functionRootPath should be /project/cloudfunctions/
    • Important: Do NOT include function name in root path
  3. Entry Point: Default is index.js with exports.main

    • Can be customized via handler parameter

Event Function Deployment

Creating New Functions:

Use manageFunctions(action="createFunction") (see MCP tool documentation for full parameter list):

  • Important: Always specify func.runtime explicitly (defaults to Nodejs18.15)
  • Provide functionRootPath as parent directory of function folders (absolute path)
  • Use force=true to overwrite existing function
  • Legacy compatibility: If an older prompt says createFunction, keep the same payload but send it through manageFunctions

Updating Function Code:

Use manageFunctions(action="updateFunctionCode"):

  • ⚠️ Note: Only updates code, cannot change runtime
  • If runtime needs to change, delete and recreate function
  • Legacy compatibility: If an older prompt says updateFunctionCode, map it to this manageFunctions action

Deployment Best Practices:

  1. Always specify runtime explicitly when creating functions
  2. Use absolute paths for functionRootPath
  3. Don't upload node_modules - dependencies installed automatically
  4. Test locally before deployment when possible
  5. Use environment variables for configuration, not hardcoded values

Core Knowledge - HTTP Functions

HTTP Function Overview

HTTP Functions are optimized for Web service scenarios, supporting standard HTTP request/response patterns.

HTTP Function coding model

  • Use a web-server style handler with req / res
  • Parse query, headers, and body from the HTTP request directly
  • Return HTTP status codes and response bodies explicitly
  • Do not use exports.main(event, context) as the primary HTTP Function shape

Key Characteristics:

  • Must listen on port 9000 (platform requirement)
  • Requires scf_bootstrap startup script
  • Supports multiple languages: Node.js, Python, Go, Java
  • Supports HTTP, SSE, WebSocket protocols

scf_bootstrap Startup Script

⚠️ CRITICAL: HTTP Functions require scf_bootstrap file

Requirement Description
File name Must be exactly scf_bootstrap (no extension)
Permission Must have execute permission (chmod +x scf_bootstrap)
Port Must start server on port 9000
Line endings Must use LF (Unix), not CRLF (Windows)

Examples:

# Node.js
#!/bin/bash
node index.js

# Python
#!/bin/bash
export PYTHONPATH="./third_party:$PYTHONPATH"
/var/lang/python310/bin/python3.10 app.py

# Go
#!/bin/bash
./main

# Java
#!/bin/bash
java -jar *.jar

HTTP Function Request & Response Handling

⚠️ CRITICAL: HTTP Functions are standard web servers. They receive full HTTP requests (method, headers, query string, body) and must return complete HTTP responses (status code, headers, body). Unlike Event Functions, there is no CloudBase-specific request envelope.

Dependencies:

  • HTTP Functions do NOT auto-install dependencies. You MUST include node_modules in the function directory or use layers.
  • Common packages: express, body-parser (built into Express 4.16+), cors, helmet

Handling Different HTTP Methods:

const express = require('express');
const app = express();

// IMPORTANT: Must parse JSON body for POST/PUT requests
app.use(express.json());

// GET - read query parameters from req.query
app.get('/users', (req, res) => {
  const { userId, page = 1, pageSize = 10 } = req.query;

  if (!userId) {
    return res.status(400).json({ error: 'userId is required' });
  }

  res.json({
    method: 'GET',
    params: { userId, page: Number(page), pageSize: Number(pageSize) },
    message: 'Query parameters processed'
  });
});

// POST - read body from req.body, query from req.query
app.post('/users', (req, res) => {
  const { action } = req.query;    // query string: ?action=create
  const { name, email } = req.body; // JSON body

  if (!name || !email) {
    return res.status(400).json({ error: 'name and email are required' });
  }

  res.json({
    method: 'POST',
    action,
    data: { name, email },
    message: 'Body and query parameters processed'
  });
});

// PUT - read custom headers from req.headers
app.put('/users/:id', (req, res) => {
  const userId = req.headers['x-user-id']; // custom header

  if (!userId) {
    return res.status(401).json({ error: 'x-user-id header is required' });
  }

  const { status } = req.body;
  res.json({
    method: 'PUT',
    userId,
    status,
    message: 'Headers and body processed'
  });
});

// For unsupported methods, return 405
app.all('*', (req, res) => {
  res.status(405).json({ error: 'Method Not Allowed' });
});

app.listen(9000); // Must be port 9000

Key Rules:

  • req.query - URL query string parameters (e.g., ?userId=123&page=1)
  • req.body - Parsed request body (requires app.use(express.json()) or equivalent middleware)
  • req.headers - All HTTP headers (lowercased, e.g., req.headers['x-user-id'])
  • req.params - URL path parameters (e.g., /users/:id)
  • Always return proper HTTP status codes: 200 (success), 400 (bad request), 401 (unauthorized), 405 (method not allowed)
  • Always call res.json(), res.send(), or res.status().json() to send a response
  • Never leave a request handler without sending a response

HTTP Function Structure & Example

my-http-function/
├── scf_bootstrap      # Required: startup script
├── package.json       # Dependencies
├── node_modules/      # Required for HTTP Functions (no auto-install)
└── index.js           # Application code

Minimal Node.js Example (Express):

const express = require('express');
const app = express();
app.use(express.json());

app.get('/', (req, res) => res.json({ message: 'Hello!' }));
app.listen(9000);  // Must be port 9000

HTTP Function Deployment

MCP Tool:

manageFunctions({
  action: "createFunction",
  func: {
    name: "myHttpFunction",
    type: "HTTP",           // Required for HTTP Function
    protocolType: "HTTP",   // "HTTP" (default) or "WS" (WebSocket)
    timeout: 60
  },
  functionRootPath: "/path/to/functions",
  force: false
})

CLI:

tcb fn deploy <name> --httpFn        # HTTP Function
tcb fn deploy <name> --httpFn --ws   # With WebSocket

⚠️ Notes:

  • Function type cannot be changed after creation
  • HTTP Functions do NOT auto-install dependencies; include node_modules or use layers

Invoking HTTP Functions

Method 1: HTTP API (with Access Token)

curl -L 'https://{envId}.api.tcloudbasegateway.com/v1/functions/{name}?webfn=true' \
  -H 'Authorization: Bearer <TOKEN>'

⚠️ Must include webfn=true parameter

Method 2: HTTP Access Service (Custom Domain)

Use manageGateway(action="createAccess") to configure HTTP access:

manageGateway({
  action: "createAccess",
  targetType: "function",
  targetName: "myHttpFunction",
  type: "HTTP",           // "HTTP" for HTTP Function
  path: "/api/hello",     // Trigger path
  // auth: false            // Optional: only set this when you intentionally need anonymous/public access
})

Notes:

  • manageFunctions(action="createFunction") 只创建 HTTP 云函数本身,不会默认替你创建 HTTP 访问路径。
  • 如果用户明确说“不需要配置 HTTP 访问服务”,不要额外调用 manageGateway(action="createAccess")。
  • 评测、浏览器或其他外部调用方可能会以匿名身份访问,而且失败后不一定会把 EXCEED_AUTHORITY 再反馈给 AI。只要场景涉及匿名 URL 访问,就要主动确认两件事:访问路径是否已创建,以及函数安全规则是否允许该访问方式。
  • 若外部请求已经报 EXCEED_AUTHORITY,先调用 readSecurityRule(resourceType="function") 查看当前规则,再根据实际需求决定是否用 writeSecurityRule(resourceType="function", aclTag="CUSTOM", rule="true") 或更细粒度规则放开权限。
# Access via default domain
curl https://{envId}.{region}.app.tcloudbase.com/{path}

# Access via custom domain
curl https://your-domain.com/{path}
Method Auth Required Use Case
HTTP API (?webfn=true) Yes (Bearer Token) Server-to-server
HTTP Access Service Optional Browser, public APIs

SSE & WebSocket Support

SSE (Server-Sent Events): Enabled by default, for server-to-client streaming (AI chat, progress updates).

// Server
res.setHeader('Content-Type', 'text/event-stream');
res.write(`data: ${JSON.stringify({ content: 'Hello' })}\n\n`);

// Client
const es = new EventSource('https://your-domain/stream');
es.onmessage = (e) => console.log(JSON.parse(e.data));

WebSocket: Enable via protocolType: "WS" in manageFunctions(action="createFunction"). For bidirectional real-time communication.

Limit Value
Idle timeout 10 - 7200 seconds
Max message size 256KB
const wss = new WebSocket.Server({ port: 9000 });
wss.on('connection', (ws) => {
  ws.on('message', (msg) => ws.send(`Echo: ${msg}`));
});

Function Logs

Querying Logs:

Primary Method: Use queryFunctions(action="listFunctionLogs") and queryFunctions(action="getFunctionLogDetail") (see MCP tool documentation).

Alternative Method (Plan B): If tools unavailable, use callCloudApi:

  1. Get Log List - Use GetFunctionLogs action:
callCloudApi({
  service: "tcb",
  action: "GetFunctionLogs",
  params: {
    EnvId: "{envId}",
    FunctionName: "functionName",
    Offset: 0,
    Limit: 10,
    StartTime: "2024-01-01 00:00:00",
    EndTime: "2024-01-01 23:59:59",
    LogRequestId: "optional-request-id",
    Qualifier: "$LATEST"
  }
})
  1. Get Log Details - Use GetFunctionLogDetail action (requires LogRequestId from step 1):
callCloudApi({
  service: "tcb",
  action: "GetFunctionLogDetail",
  params: {
    StartTime: "2024-01-01 00:00:00",
    EndTime: "2024-01-01 23:59:59",
    LogRequestId: "request-id-from-log-list"
  }
})

Log Query Limitations:

  • Offset + Limit cannot exceed 10000
  • StartTime and EndTime interval cannot exceed 1 day
  • Use pagination for large time ranges

Log Query Best Practices:

  1. Query logs within 1-day windows
  2. Use RequestId for specific invocation debugging
  3. Combine list and detail queries for comprehensive debugging
  4. Check logs after deployment to verify function behavior

Invoking Event Functions

From Web Applications:

import cloudbaseSDK from "@cloudbase/js-sdk";

const cloudbase = cloudbaseSDK.init({
  env: 'your-env-id',
  region: 'ap-shanghai',
  accessKey: 'your-access-key'
});

// Call event function
const result = await cloudbase.callFunction({
  name: "functionName",
  data: { /* function parameters */ }
});

From Mini Programs:

wx.cloud.callFunction({
  name: "functionName",
  data: { /* function parameters */ }
}).then(res => {
  console.log(res.result);
});

From Node.js Backend:

const cloudbase = require("@cloudbase/node-sdk");

const app = cloudbase.init({
  env: "your-env-id"
});

const result = await app.callFunction({
  name: "functionName",
  data: { /* function parameters */ }
});

From HTTP API:

Use CloudBase HTTP API to invoke event functions:

  • Endpoint: https://{envId}.api.tcloudbasegateway.com/v1/functions/{functionName}
  • Requires authentication token (Bearer Token)
  • See http-api skill for details

HTTP Access Configuration (for Event Functions)

HTTP Access vs HTTP API:

  • HTTP API: Uses CloudBase API endpoint with authentication token
  • HTTP Access: Creates direct HTTP/HTTPS endpoint for standard REST API access without SDK
  • Important: This section applies to Event Functions only. It does not change an HTTP Function into an Event Function or vice versa.

Creating HTTP Access:

Primary Method: Use manageGateway(action="createAccess") (see MCP tool documentation).

Alternative Method (Plan B): If tool unavailable, use callCloudApi with CreateCloudBaseGWAPI:

callCloudApi({
  service: "tcb",
  action: "CreateCloudBaseGWAPI",
  params: {
    EnableUnion: true,
    Path: "/api/users",
    ServiceId: "{envId}",
    Type: 6,
    Name: "functionName",
    AuthSwitch: 2,
    PathTransmission: 2,
    EnableRegion: true,
    Domain: "*"  // Use "*" for default domain, or custom domain name
  }
})

Key Parameters:

  • Type: 6 - Cloud Function type (required)
  • AuthSwitch: 2 - No auth (1 = with auth)
  • Domain: "*" - Default domain, or specify custom domain

Access URL: https://{envId}.{region}.app.tcloudbase.com/{path} or https://{domain}/{path}

Function Configuration

Environment Variables:

Set via func.envVariables when creating/updating:

{
  envVariables: {
    "DATABASE_URL": "mysql://...",
    "API_KEY": "secret-key"
  }
}

⚠️ CRITICAL: Environment Variable Update Constraint

When updating environment variables for existing functions:

  1. MUST first query current environment variables using queryFunctions(action="getFunctionDetail") to get the function's current configuration
  2. MUST merge new environment variables with existing ones
  3. DO NOT directly overwrite - this will delete existing environment variables not included in the update

Correct Update Pattern:

// 1. First, get current function details
const currentFunction = await queryFunctions({
  action: "getFunctionDetail",
  functionName: "functionName"
});

// 2. Merge existing envVariables with new ones
const mergedEnvVariables = {
  ...currentFunction.EnvVariables,  // Existing variables
  ...newEnvVariables                 // New/updated variables
};

// 3. Update with merged variables
await manageFunctions({
  action: "updateFunctionConfig",
  functionName: "functionName",
  envVariables: mergedEnvVariables
});

Why This Matters:

  • Direct overwrite will delete all environment variables not included in the update
  • This can break function functionality if critical variables are removed
  • Always preserve existing configuration when making partial updates

Timeout Configuration:

Set via func.timeout (in seconds):

  • Default timeout varies by runtime
  • Maximum timeout depends on runtime version
  • Consider function execution time when setting

Timer Triggers:

Configure via func.triggers:

  • Type: timer (only supported type)
  • Config: Cron expression (7 fields: second minute hour day month week year)
  • Examples:
    • "0 0 2 1 * * *" - 2:00 AM on 1st of every month
    • "0 30 9 * * * *" - 9:30 AM every day

VPC Configuration:

For accessing VPC resources:

{
  vpc: {
    vpcId: "vpc-xxxxx",
    subnetId: "subnet-xxxxx"
  }
}

MCP Tools Reference

Function Management:

  • queryFunctions(action="listFunctions"|"getFunctionDetail") - Preferred read entrance for function inventory and detail
  • manageFunctions(action="createFunction") - Create cloud function (supports both Event and HTTP types via func.type)
    • func.type: "Event" - Event Function (default)
    • func.type: "HTTP" - HTTP Function
    • func.protocolType: "WS" - Enable WebSocket for HTTP Function
  • manageFunctions(action="updateFunctionCode") - Update function code (runtime cannot change)
  • manageFunctions(action="updateFunctionConfig") - Update function configuration (⚠️ when updating envVariables, must first query and merge with existing values to avoid overwriting)
  • Legacy aliases still seen in historical prompts: getFunctionList, createFunction, updateFunctionCode, updateFunctionConfig

Logging:

  • queryFunctions(action="listFunctionLogs") - Get function log list (basic info)
  • queryFunctions(action="getFunctionLogDetail") - Get detailed log content by RequestId
  • callCloudApi (Plan B) - Use GetFunctionLogs and GetFunctionLogDetail actions if direct tools unavailable
  • Legacy aliases still seen in historical prompts: getFunctionLogs, getFunctionLogDetail

HTTP Access:

  • queryGateway(action="getAccess") - Query current function gateway exposure
  • manageGateway(action="createAccess") - Create HTTP access for function
  • callCloudApi (Plan B) - Use CreateCloudBaseGWAPI action if direct tool unavailable
  • Legacy alias still seen in historical prompts: createFunctionHTTPAccess

Triggers:

  • queryFunctions(action="listFunctionTriggers") - Inspect triggers
  • manageFunctions(action="createFunctionTrigger"|"deleteFunctionTrigger") - Create or delete function triggers
  • Legacy alias still seen in historical prompts: manageFunctionTriggers

CLI Commands:

  • tcb fn deploy <name> - Deploy Event Function
  • tcb fn deploy <name> --httpFn - Deploy HTTP Function
  • tcb fn deploy <name> --httpFn --ws - Deploy HTTP Function with WebSocket
  • tcb fn deploy --all - Deploy all functions in config

Common Patterns

Error Handling

exports.main = async (event, context) => {
  try {
    // Function logic
    return {
      code: 0,
      message: "Success",
      data: result
    };
  } catch (error) {
    return {
      code: -1,
      message: error.message,
      data: null
    };
  }
};

Environment Variable Usage

exports.main = async (event, context) => {
  const apiKey = process.env.API_KEY;
  const dbUrl = process.env.DATABASE_URL;
  
  // Use environment variables
};

Database Operations

const cloudbase = require("@cloudbase/node-sdk");

const app = cloudbase.init({
  env: process.env.ENV_ID
});

exports.main = async (event, context) => {
  const db = app.database();
  const result = await db.collection("users").get();
  return result;
};

Best Practices

General Best Practices

  1. Runtime Selection: Always specify runtime explicitly, use Nodejs18.15 for new projects
  2. Code Organization: Keep functions focused and single-purpose
  3. Error Handling: Always implement proper error handling
  4. Environment Variables: Use env vars for configuration, never hardcode secrets
  5. Logging: Add meaningful logs for debugging
  6. Testing: Test functions locally when possible before deployment
  7. Security: Implement authentication/authorization for HTTP access
  8. Performance: Optimize cold start time, use connection pooling for databases
  9. Monitoring: Regularly check logs and monitor function performance
  10. Documentation: Document function parameters and return values

HTTP Function Specific Best Practices

  1. Port Configuration: Always listen on port 9000
  2. scf_bootstrap: Ensure correct file permissions and LF line endings
  3. Health Check: Add /health endpoint for monitoring
  4. CORS: Configure CORS headers for browser access
  5. Graceful Shutdown: Handle process signals properly
  6. Dependencies: Include node_modules in package or use layers (no auto-install for HTTP Functions)
  7. Timeout: Set appropriate timeout for long-running SSE/WebSocket connections
  8. Error Responses: Return proper HTTP status codes and error messages

Choosing Between Event and HTTP Functions

Scenario Recommended Type
SDK/Mini Program calls Event Function
Scheduled tasks (cron) Event Function
REST API / Web services HTTP Function
SSE streaming (AI chat) HTTP Function
WebSocket real-time HTTP Function
File upload/download HTTP Function
Multi-language support HTTP Function

Related Skills

  • cloudrun-development - For container-based backend services
  • http-api - For HTTP API invocation patterns
  • cloudbase-platform - For general CloudBase platform knowledge

04 同仓库更多 Skills

YYC-Cube
YYC-Cube/YYC3-Knowledge-Base

vercel-sandbox

Run agent-browser + Chrome inside Vercel Sandbox microVMs for browser automation from any Vercel-deployed app. Use when the user needs browser automation in a Vercel app (Next.js, SvelteKit, Nuxt, Remix, Astro, etc.), wants to run headless Chrome without binary size limits, needs persistent browser sessions across commands, or wants ephemeral isolated browser environments. Triggers include "Vercel Sandbox browser", "microVM Chrome", "agent-browser in sandbox", "browser automation on Vercel", or any task requiring Chrome in a Vercel Sandbox.

★ 1
YYC-Cube
YYC-Cube/YYC3-Knowledge-Base

slack

Interact with Slack workspaces using browser automation. Use when the user needs to check unread channels, navigate Slack, send messages, extract data, find information, search conversations, or automate any Slack task. Triggers include "check my Slack", "what channels have unreads", "send a message to", "search Slack for", "extract from Slack", "find who said", or any task requiring programmatic Slack interaction.

★ 1
YYC-Cube
YYC-Cube/YYC3-Knowledge-Base

electron

Automate Electron desktop apps (VS Code, Slack, Discord, Figma, Notion, Spotify, etc.) using agent-browser via Chrome DevTools Protocol. Use when the user needs to interact with an Electron app, automate a desktop app, connect to a running app, control a native app, or test an Electron application. Triggers include "automate Slack app", "control VS Code", "interact with Discord app", "test this Electron app", "connect to desktop app", or any task requiring automation of a native Electron application.

★ 1
YYC-Cube
YYC-Cube/YYC3-Knowledge-Base

dogfood

Systematically explore and test a web application to find bugs, UX issues, and other problems. Use when asked to "dogfood", "QA", "exploratory test", "find issues", "bug hunt", "test this app/site/platform", or review the quality of a web application. Produces a structured report with full reproduction evidence -- step-by-step screenshots, repro videos, and detailed repro steps for every issue -- so findings can be handed directly to the responsible teams.

★ 1
YYC-Cube
YYC-Cube/YYC3-Knowledge-Base

agent-browser

Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, or automating any browser task. Triggers include requests to "open a website", "fill out a form", "click a button", "take a screenshot", "scrape data from a page", "test this web app", "login to a site", "automate browser actions", or any task requiring programmatic web interaction.

★ 1
YYC-Cube
YYC-Cube/YYC3-Knowledge-Base

shadcn

Manages shadcn components and projects — adding, searching, fixing, debugging, styling, and composing UI, including chat interfaces. Provides project context, component docs, and usage examples. Applies when working with shadcn/ui, component registries, presets, --preset codes, or any project with a components.json file. Also triggers for "shadcn init", "create an app with --preset", "switch to --preset", and for building a design system or applying a DESIGN.md or brand spec to shadcn/ui.

★ 1