重要前提
安装AI Skills的关键前提是:必须科学上网,且开启TUN模式,这一点至关重要,直接决定安装能否顺利完成,在此郑重提醒三遍:科学上网,科学上网,科学上网。查看完整安装教程 →
hyperforce-2025 by josiahsiegel/claude-plugin-marketplace
npx skills add https://github.com/josiahsiegel/claude-plugin-marketplace --skill hyperforce-2025强制要求:在 Windows 上始终对文件路径使用反斜杠
在 Windows 上使用编辑或写入工具时,必须在文件路径中使用反斜杠 (\),而不是正斜杠 (/)。
示例:
D:/repos/project/file.tsxD:\repos\project\file.tsx这适用于:
除非用户明确要求,否则切勿创建新的文档文件。
广告位招租
在这里展示您的产品或服务
触达数万 AI 开发者,精准高效
Hyperforce 是 Salesforce 基于公共云平台(AWS、Azure、Google Cloud)构建的下一代基础设施架构。它代表了 Salesforce 从基于数据中心的基础设施到云原生、容器化微服务的完全重新架构。
关键创新:作为代码的基础设施,可以部署在任何地方,为客户提供选择、控制和数据驻留合规性。
传统方式:修补和更新现有服务器 Hyperforce:每次部署时销毁并重新创建服务器
Old Architecture:
Server → Patch → Patch → Patch → Configuration Drift
Hyperforce:
Container Image v1 → Deploy
New Code → Build Container Image v2 → Replace v1 with v2
Result: Every deployment is identical, reproducible
优势:
架构:
Region: US-East (Virginia)
├─ Availability Zone A (Data Center 1)
│ ├─ App Servers (Kubernetes pods)
│ ├─ Database Primary
│ └─ Load Balancer
├─ Availability Zone B (Data Center 2)
│ ├─ App Servers (Kubernetes pods)
│ ├─ Database Replica
│ └─ Load Balancer
└─ Availability Zone C (Data Center 3)
├─ App Servers (Kubernetes pods)
├─ Database Replica
└─ Load Balancer
Traffic Distribution: Round-robin across all AZs
Failure Handling: If AZ fails, traffic routes to remaining AZs
RTO (Recovery Time Objective): <5 minutes
RPO (Recovery Point Objective): <30 seconds
对开发者的影响:
传统方式:边界安全(防火墙保护内部所有内容) Hyperforce:无隐式信任 - 始终验证一切
Zero Trust Model:
├─ Identity Verification (MFA required for all users by 2025)
├─ Device Trust (managed devices only)
├─ Network Segmentation (micro-segmentation between services)
├─ Least Privilege Access (minimal permissions by default)
├─ Continuous Monitoring (real-time threat detection)
└─ Encryption Everywhere (TLS 1.3, data at rest encryption)
代码影响:
// OLD: Assume internal traffic is safe
public without sharing class InternalService {
// No auth checks - trusted network
}
// HYPERFORCE: Always verify, never trust
public with sharing class InternalService {
// Always enforce sharing rules
// Always validate session
// Always check field-level security
public List<Account> getAccounts() {
// WITH SECURITY_ENFORCED prevents data leaks
return [SELECT Id, Name FROM Account WITH SECURITY_ENFORCED];
}
}
2025 年要求:
所有内容都定义为代码,并进行版本控制:
# Hyperforce deployment manifest (conceptual)
apiVersion: hyperforce.salesforce.com/v1
kind: SalesforceOrg
metadata:
name: production-org
region: aws-us-east-1
spec:
edition: enterprise
features:
- agentforce
- dataCloud
- einstein
compute:
pods: 50
autoScaling:
min: 10
max: 100
targetCPU: 70%
storage:
size: 500GB
replication: 3
backup:
frequency: hourly
retention: 30days
networking:
privateLink: enabled
ipWhitelist:
- 203.0.113.0/24
对开发者的优势:
Hyperforce 从头开始重建:
┌────────────────────────────────────────────────────────┐
│ AWS Region (us-east-1) │
├────────────────────────────────────────────────────────┤
│ VPC (Virtual Private Cloud) │
│ ├─ Public Subnets (3 AZs) │
│ │ └─ Application Load Balancer (ALB) │
│ ├─ Private Subnets (3 AZs) │
│ │ ├─ EKS Cluster (Kubernetes) │
│ │ │ ├─ Salesforce App Pods (autoscaling) │
│ │ │ ├─ Metadata Service Pods │
│ │ │ ├─ API Gateway Pods │
│ │ │ └─ Background Job Pods (Batch, Scheduled) │
│ │ ├─ RDS Aurora PostgreSQL (multi-AZ) │
│ │ ├─ ElastiCache Redis (session storage) │
│ │ └─ S3 Buckets (attachments, documents) │
│ └─ Database Subnets (3 AZs) │
│ └─ Aurora Database Cluster │
├────────────────────────────────────────────────────────┤
│ Additional Services │
│ ├─ CloudWatch (monitoring, logs) │
│ ├─ CloudTrail (audit logs) │
│ ├─ AWS Shield (DDoS protection) │
│ ├─ AWS WAF (web application firewall) │
│ ├─ KMS (encryption key management) │
│ └─ PrivateLink (secure connectivity) │
└────────────────────────────────────────────────────────┘
使用的 AWS 服务:
Azure Region (East US)
├─ Virtual Network (VNet)
│ ├─ AKS (Azure Kubernetes Service)
│ │ └─ Salesforce workloads
│ ├─ Azure Database for PostgreSQL (Hyperscale)
│ ├─ Azure Cache for Redis
│ └─ Azure Blob Storage
├─ Azure Front Door (CDN + Load Balancer)
├─ Azure Monitor (logging, metrics)
├─ Azure Active Directory (identity)
└─ Azure Key Vault (secrets, encryption)
GCP Region (us-central1)
├─ VPC Network
│ ├─ GKE (Google Kubernetes Engine)
│ ├─ Cloud SQL (PostgreSQL)
│ ├─ Memorystore (Redis)
│ └─ Cloud Storage (GCS)
├─ Cloud Load Balancing
├─ Cloud Armor (DDoS protection)
├─ Cloud Monitoring (Stackdriver)
└─ Cloud KMS (encryption)
可用的 Hyperforce 区域:
Americas:
├─ US East (Virginia) - AWS, Azure
├─ US West (Oregon) - AWS
├─ US Central (Iowa) - GCP
├─ Canada (Toronto) - AWS
└─ Brazil (São Paulo) - AWS
Europe:
├─ UK (London) - AWS
├─ Germany (Frankfurt) - AWS, Azure
├─ France (Paris) - AWS
├─ Ireland (Dublin) - AWS
└─ Switzerland (Zurich) - AWS
Asia Pacific:
├─ Japan (Tokyo) - AWS
├─ Australia (Sydney) - AWS
├─ Singapore - AWS
├─ India (Mumbai) - AWS
└─ South Korea (Seoul) - AWS
Middle East:
└─ UAE (Dubai) - AWS
保留在区域内的内容:
可能离开区域的内容:
代码影响:
// Data residency automatically enforced
// No code changes needed - Hyperforce handles it
// Example: File stored in org's region
ContentVersion cv = new ContentVersion(
Title = 'Customer Contract',
PathOnClient = 'contract.pdf',
VersionData = Blob.valueOf('contract data')
);
insert cv;
// File automatically stored in:
// - AWS S3 in org's region
// - Encrypted at rest (AES-256)
// - Replicated across 3 AZs in region
// - Never leaves region boundary
Hyperforce 维护:
旧架构(基于数据中心):
User (Germany) → Transatlantic cable → US Data Center → Response
Latency: 150-200ms
Hyperforce:
User (Germany) → Frankfurt Hyperforce Region → Response
Latency: 10-30ms
Result: 5-10x faster for regional users
传统方式:固定容量,必须为峰值负载预置资源 Hyperforce:基于需求的动态扩展
Business Hours (9 AM - 5 PM):
├─ High user load
├─ Kubernetes scales up pods: 50 → 150
└─ Response times maintained
Off Hours (6 PM - 8 AM):
├─ Low user load
├─ Kubernetes scales down pods: 150 → 30
└─ Cost savings (pay for what you use)
Black Friday (peak event):
├─ Extreme load
├─ Kubernetes scales to maximum: 30 → 500 pods in minutes
└─ No downtime, no performance degradation
限制 - 无变化:
// Hyperforce does NOT change governor limits
// Limits remain the same as classic Salesforce:
// - 100 SOQL queries per transaction
// - 150 DML statements
// - 6 MB heap size (sync), 12 MB (async)
// But: Infrastructure scales to handle more concurrent users
Salesforce 处理迁移(无需客户操作):
Phase 1: Assessment (Salesforce internal)
├─ Analyze org size, customizations
├─ Identify any incompatible features
└─ Plan migration window
Phase 2: Pre-Migration (Customer notified)
├─ Salesforce sends notification (90 days notice)
├─ Customer tests in sandbox (migrated first)
└─ Customer validates functionality
Phase 3: Migration (Weekend maintenance window)
├─ Backup all data
├─ Replicate data to Hyperforce
├─ Cutover DNS (redirect traffic)
└─ Validate migration success
Phase 4: Post-Migration
├─ Monitor performance
├─ Support customer issues
└─ Decommission old infrastructure
Downtime: Typically <2 hours
无需代码更改:
// Your Apex code works identically on Hyperforce
public class MyController {
public List<Account> getAccounts() {
return [SELECT Id, Name FROM Account LIMIT 10];
}
}
// No changes needed
// Same APIs, same limits, same behavior
潜在的性能改进:
向后兼容性:与现有代码 100% 兼容
使用沙盒迁移:
1. Salesforce migrates your sandbox first
2. Test all critical functionality:
├─ Custom Apex classes
├─ Triggers and workflows
├─ Integrations (API callouts)
├─ Lightning components
└─ Reports and dashboards
3. Validate performance:
├─ Run load tests
├─ Check API response times
└─ Verify batch jobs complete
4. Report any issues to Salesforce
5. Production migration scheduled after sandbox validated
Hyperforce 公开基础设施 API:
// Query org's Hyperforce region (API 62.0+)
Organization org = [SELECT Id, InstanceName, InfrastructureRegion__c FROM Organization LIMIT 1];
System.debug('Region: ' + org.InfrastructureRegion__c); // 'aws-us-east-1'
// Check if org is on Hyperforce
System.debug('Is Hyperforce: ' + org.IsHyperforce__c); // true
AWS PrivateLink / Azure Private Link:
Traditional: Salesforce API → Public Internet → Your API
Security: TLS encryption, but still public internet
Hyperforce PrivateLink: Salesforce API → Private Network → Your API
Security: Never touches public internet, lower latency
Setup:
1. Create VPC Endpoint (AWS) or Private Endpoint (Azure)
2. Salesforce provides service endpoint name
3. Configure Named Credential in Salesforce with private endpoint
4. API calls route over private network
配置:
// Named Credential uses PrivateLink endpoint
// Setup → Named Credentials → External API (PrivateLink)
// URL: https://api.internal.example.com (private endpoint)
// Apex callout
HttpRequest req = new HttpRequest();
req.setEndpoint('callout:ExternalAPIPrivateLink/data');
req.setMethod('GET');
Http http = new Http();
HttpResponse res = http.send(req);
// Callout never leaves private network
// Lower latency, higher security
CloudWatch / Azure Monitor 集成:
Salesforce publishes metrics to your cloud account:
├─ API request volume
├─ API response times
├─ Error rates
├─ Governor limit usage
└─ Batch job completion times
Benefits:
- Unified monitoring (Salesforce + your apps)
- Custom alerting (CloudWatch Alarms)
- Cost attribution (AWS Cost Explorer)
预期增强功能:
Hyperforce 代表了 Salesforce 对现代化、云原生基础设施的承诺,该基础设施可在全球范围内扩展,同时满足最严格的合规性和性能要求。
每周安装次数
62
仓库
GitHub 星标数
21
首次出现
2026年1月24日
安全审计
已安装于
claude-code48
opencode47
gemini-cli46
codex44
cursor42
antigravity39
MANDATORY: Always Use Backslashes on Windows for File Paths
When using Edit or Write tools on Windows, you MUST use backslashes (\) in file paths, NOT forward slashes (/).
Examples:
D:/repos/project/file.tsxD:\repos\project\file.tsxThis applies to:
NEVER create new documentation files unless explicitly requested by the user.
Hyperforce is Salesforce's next-generation infrastructure architecture built on public cloud platforms (AWS, Azure, Google Cloud). It represents a complete re-architecture of Salesforce from data center-based infrastructure to cloud-native, containerized microservices.
Key Innovation : Infrastructure as code that can be deployed anywhere, giving customers choice, control, and data residency compliance.
Traditional : Patch and update existing servers Hyperforce : Destroy and recreate servers with each deployment
Old Architecture:
Server → Patch → Patch → Patch → Configuration Drift
Hyperforce:
Container Image v1 → Deploy
New Code → Build Container Image v2 → Replace v1 with v2
Result: Every deployment is identical, reproducible
Benefits :
Architecture :
Region: US-East (Virginia)
├─ Availability Zone A (Data Center 1)
│ ├─ App Servers (Kubernetes pods)
│ ├─ Database Primary
│ └─ Load Balancer
├─ Availability Zone B (Data Center 2)
│ ├─ App Servers (Kubernetes pods)
│ ├─ Database Replica
│ └─ Load Balancer
└─ Availability Zone C (Data Center 3)
├─ App Servers (Kubernetes pods)
├─ Database Replica
└─ Load Balancer
Traffic Distribution: Round-robin across all AZs
Failure Handling: If AZ fails, traffic routes to remaining AZs
RTO (Recovery Time Objective): <5 minutes
RPO (Recovery Point Objective): <30 seconds
Impact on Developers :
Traditional : Perimeter security (firewall protects everything inside) Hyperforce : No implicit trust - verify everything, always
Zero Trust Model:
├─ Identity Verification (MFA required for all users by 2025)
├─ Device Trust (managed devices only)
├─ Network Segmentation (micro-segmentation between services)
├─ Least Privilege Access (minimal permissions by default)
├─ Continuous Monitoring (real-time threat detection)
└─ Encryption Everywhere (TLS 1.3, data at rest encryption)
Code Impact :
// OLD: Assume internal traffic is safe
public without sharing class InternalService {
// No auth checks - trusted network
}
// HYPERFORCE: Always verify, never trust
public with sharing class InternalService {
// Always enforce sharing rules
// Always validate session
// Always check field-level security
public List<Account> getAccounts() {
// WITH SECURITY_ENFORCED prevents data leaks
return [SELECT Id, Name FROM Account WITH SECURITY_ENFORCED];
}
}
2025 Requirements :
Everything defined as code, version-controlled :
# Hyperforce deployment manifest (conceptual)
apiVersion: hyperforce.salesforce.com/v1
kind: SalesforceOrg
metadata:
name: production-org
region: aws-us-east-1
spec:
edition: enterprise
features:
- agentforce
- dataCloud
- einstein
compute:
pods: 50
autoScaling:
min: 10
max: 100
targetCPU: 70%
storage:
size: 500GB
replication: 3
backup:
frequency: hourly
retention: 30days
networking:
privateLink: enabled
ipWhitelist:
- 203.0.113.0/24
Benefits for Developers :
Hyperforce rebuilt from scratch :
┌────────────────────────────────────────────────────────┐
│ AWS Region (us-east-1) │
├────────────────────────────────────────────────────────┤
│ VPC (Virtual Private Cloud) │
│ ├─ Public Subnets (3 AZs) │
│ │ └─ Application Load Balancer (ALB) │
│ ├─ Private Subnets (3 AZs) │
│ │ ├─ EKS Cluster (Kubernetes) │
│ │ │ ├─ Salesforce App Pods (autoscaling) │
│ │ │ ├─ Metadata Service Pods │
│ │ │ ├─ API Gateway Pods │
│ │ │ └─ Background Job Pods (Batch, Scheduled) │
│ │ ├─ RDS Aurora PostgreSQL (multi-AZ) │
│ │ ├─ ElastiCache Redis (session storage) │
│ │ └─ S3 Buckets (attachments, documents) │
│ └─ Database Subnets (3 AZs) │
│ └─ Aurora Database Cluster │
├────────────────────────────────────────────────────────┤
│ Additional Services │
│ ├─ CloudWatch (monitoring, logs) │
│ ├─ CloudTrail (audit logs) │
│ ├─ AWS Shield (DDoS protection) │
│ ├─ AWS WAF (web application firewall) │
│ ├─ KMS (encryption key management) │
│ └─ PrivateLink (secure connectivity) │
└────────────────────────────────────────────────────────┘
AWS Services Used :
Azure Region (East US)
├─ Virtual Network (VNet)
│ ├─ AKS (Azure Kubernetes Service)
│ │ └─ Salesforce workloads
│ ├─ Azure Database for PostgreSQL (Hyperscale)
│ ├─ Azure Cache for Redis
│ └─ Azure Blob Storage
├─ Azure Front Door (CDN + Load Balancer)
├─ Azure Monitor (logging, metrics)
├─ Azure Active Directory (identity)
└─ Azure Key Vault (secrets, encryption)
GCP Region (us-central1)
├─ VPC Network
│ ├─ GKE (Google Kubernetes Engine)
│ ├─ Cloud SQL (PostgreSQL)
│ ├─ Memorystore (Redis)
│ └─ Cloud Storage (GCS)
├─ Cloud Load Balancing
├─ Cloud Armor (DDoS protection)
├─ Cloud Monitoring (Stackdriver)
└─ Cloud KMS (encryption)
Available Hyperforce Regions :
Americas:
├─ US East (Virginia) - AWS, Azure
├─ US West (Oregon) - AWS
├─ US Central (Iowa) - GCP
├─ Canada (Toronto) - AWS
└─ Brazil (São Paulo) - AWS
Europe:
├─ UK (London) - AWS
├─ Germany (Frankfurt) - AWS, Azure
├─ France (Paris) - AWS
├─ Ireland (Dublin) - AWS
└─ Switzerland (Zurich) - AWS
Asia Pacific:
├─ Japan (Tokyo) - AWS
├─ Australia (Sydney) - AWS
├─ Singapore - AWS
├─ India (Mumbai) - AWS
└─ South Korea (Seoul) - AWS
Middle East:
└─ UAE (Dubai) - AWS
What stays in region :
What may leave region :
Code Implication :
// Data residency automatically enforced
// No code changes needed - Hyperforce handles it
// Example: File stored in org's region
ContentVersion cv = new ContentVersion(
Title = 'Customer Contract',
PathOnClient = 'contract.pdf',
VersionData = Blob.valueOf('contract data')
);
insert cv;
// File automatically stored in:
// - AWS S3 in org's region
// - Encrypted at rest (AES-256)
// - Replicated across 3 AZs in region
// - Never leaves region boundary
Hyperforce maintains :
Old Architecture (data center-based):
User (Germany) → Transatlantic cable → US Data Center → Response
Latency: 150-200ms
Hyperforce :
User (Germany) → Frankfurt Hyperforce Region → Response
Latency: 10-30ms
Result: 5-10x faster for regional users
Traditional : Fixed capacity, must provision for peak load Hyperforce : Dynamic scaling based on demand
Business Hours (9 AM - 5 PM):
├─ High user load
├─ Kubernetes scales up pods: 50 → 150
└─ Response times maintained
Off Hours (6 PM - 8 AM):
├─ Low user load
├─ Kubernetes scales down pods: 150 → 30
└─ Cost savings (pay for what you use)
Black Friday (peak event):
├─ Extreme load
├─ Kubernetes scales to maximum: 30 → 500 pods in minutes
└─ No downtime, no performance degradation
Governor Limits - No Change :
// Hyperforce does NOT change governor limits
// Limits remain the same as classic Salesforce:
// - 100 SOQL queries per transaction
// - 150 DML statements
// - 6 MB heap size (sync), 12 MB (async)
// But: Infrastructure scales to handle more concurrent users
Salesforce handles migration (no customer action required):
Phase 1: Assessment (Salesforce internal)
├─ Analyze org size, customizations
├─ Identify any incompatible features
└─ Plan migration window
Phase 2: Pre-Migration (Customer notified)
├─ Salesforce sends notification (90 days notice)
├─ Customer tests in sandbox (migrated first)
└─ Customer validates functionality
Phase 3: Migration (Weekend maintenance window)
├─ Backup all data
├─ Replicate data to Hyperforce
├─ Cutover DNS (redirect traffic)
└─ Validate migration success
Phase 4: Post-Migration
├─ Monitor performance
├─ Support customer issues
└─ Decommission old infrastructure
Downtime: Typically <2 hours
No Code Changes Required :
// Your Apex code works identically on Hyperforce
public class MyController {
public List<Account> getAccounts() {
return [SELECT Id, Name FROM Account LIMIT 10];
}
}
// No changes needed
// Same APIs, same limits, same behavior
Potential Performance Improvements :
Backward Compatibility : 100% compatible with existing code
Use Sandbox Migration :
1. Salesforce migrates your sandbox first
2. Test all critical functionality:
├─ Custom Apex classes
├─ Triggers and workflows
├─ Integrations (API callouts)
├─ Lightning components
└─ Reports and dashboards
3. Validate performance:
├─ Run load tests
├─ Check API response times
└─ Verify batch jobs complete
4. Report any issues to Salesforce
5. Production migration scheduled after sandbox validated
Hyperforce exposes infrastructure APIs :
// Query org's Hyperforce region (API 62.0+)
Organization org = [SELECT Id, InstanceName, InfrastructureRegion__c FROM Organization LIMIT 1];
System.debug('Region: ' + org.InfrastructureRegion__c); // 'aws-us-east-1'
// Check if org is on Hyperforce
System.debug('Is Hyperforce: ' + org.IsHyperforce__c); // true
AWS PrivateLink / Azure Private Link :
Traditional: Salesforce API → Public Internet → Your API
Security: TLS encryption, but still public internet
Hyperforce PrivateLink: Salesforce API → Private Network → Your API
Security: Never touches public internet, lower latency
Setup:
1. Create VPC Endpoint (AWS) or Private Endpoint (Azure)
2. Salesforce provides service endpoint name
3. Configure Named Credential in Salesforce with private endpoint
4. API calls route over private network
Configuration :
// Named Credential uses PrivateLink endpoint
// Setup → Named Credentials → External API (PrivateLink)
// URL: https://api.internal.example.com (private endpoint)
// Apex callout
HttpRequest req = new HttpRequest();
req.setEndpoint('callout:ExternalAPIPrivateLink/data');
req.setMethod('GET');
Http http = new Http();
HttpResponse res = http.send(req);
// Callout never leaves private network
// Lower latency, higher security
CloudWatch / Azure Monitor Integration :
Salesforce publishes metrics to your cloud account:
├─ API request volume
├─ API response times
├─ Error rates
├─ Governor limit usage
└─ Batch job completion times
Benefits:
- Unified monitoring (Salesforce + your apps)
- Custom alerting (CloudWatch Alarms)
- Cost attribution (AWS Cost Explorer)
Expected Enhancements :
Hyperforce represents Salesforce's commitment to modern, cloud-native infrastructure that scales globally while meeting the most stringent compliance and performance requirements.
Weekly Installs
62
Repository
GitHub Stars
21
First Seen
Jan 24, 2026
Security Audits
Gen Agent Trust HubPassSocketPassSnykPass
Installed on
claude-code48
opencode47
gemini-cli46
codex44
cursor42
antigravity39
Supabase Postgres 最佳实践指南 - 8大类别性能优化规则与SQL示例
81,400 周安装
Ruby on Rails开发最佳实践指南:45条规则优化性能与可维护性
61 周安装
Next.js开发技能:构建现代全栈Web应用指南(App Router/服务器组件/SEO优化)
76 周安装
RWKV架构详解:融合Transformer与RNN优势的高效AI模型安装与使用指南
99 周安装
Outlines:结构化文本生成库,保证JSON/XML/代码输出,支持Pydantic与本地模型
99 周安装
nnsight远程可解释性工具:透明访问PyTorch模型内部,支持本地与远程大模型
99 周安装
NeMo Curator - NVIDIA GPU 加速的大语言模型数据整理工具包,去重快16倍
104 周安装