From high-perf-backend
High-performance backend development covering Java (Spring Boot 3, Microservices), Rust (Axum/Actix-web/Tokio), and Go (Gin, gRPC, stdlib). Use whenever the user builds, optimizes, or architects backend services or APIs in these languages. Triggers: "backend", "API server", "microservice", "high concurrency", "low latency", "Spring Boot", "Axum", "Actix", "Tokio", "Gin", "gRPC", "circuit breaker", "rate limiting", "connection pool", "cache strategy", "distributed system", "observability", "profiling", "benchmarking", or any backend design/review/optimization in Java, Rust, or Go. Also trigger for code review, DB schema design for high-load, CI/CD for microservices, or general "help me write a REST API" in these stacks.
How this skill is triggered — by the user, by Claude, or both
Slash command
/high-perf-backend:high-perf-backendThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
This skill provides battle-tested patterns, architecture guidance, and implementation best practices
This skill provides battle-tested patterns, architecture guidance, and implementation best practices for building production-grade, high-performance backend services in Java, Rust, and Go.
It includes a 技术选型决策矩阵 (tech stack decision matrix) covering 40+ real-world business scenarios, and an interactive workflow to help users land on the right technology combination for their specific needs.
Before making any technology recommendation, collect the following information from the user. Use clear, structured questions. If the user has already provided some of this info in their message, skip those questions and confirm what you understood.
Required questions (ask all that are not already answered):
业务场景 (Business Scenario)
规模与性能要求 (Scale & Performance)
技术偏好与约束 (Tech Preferences & Constraints)
特殊需求 (Special Requirements)
After gathering user input, search the tech stack decision matrix CSV for matching scenarios.
Step 1: Read the CSV
Read references/tech-stack-matrix.csv and parse it.
Step 2: Score and rank rows
For each row in the CSV, calculate a relevance score based on how many user-provided keywords
appear in these columns: scenario, tags, notes, scale, latency_req, consistency_req, language.
Scoring rules:
tags column: +3 pointsscenario or notes: +2 pointsscale level: +2 pointslatency_req: +2 pointsconsistency_req: +1 pointlanguage (if user specified): +3 pointsStep 3: Select top matches Take the top 1-3 rows by score. These are the reference architectures to present to the user.
Based on the matched rows, synthesize a technology recommendation structured as follows:
## 推荐技术方案 (Recommended Tech Stack)
### 方案概述 (Overview)
- 参考场景: [matched scenario name(s)]
- 推荐语言/框架: [language + framework]
- 核心理由: [why this combination fits]
### 技术选型详情 (Stack Details)
| 层级 | 选择 | 理由 |
|------|------|------|
| 语言/框架 | ... | ... |
| 通信协议 | ... | ... |
| 主数据库 | ... | ... |
| 辅助存储 | ... | ... |
| 缓存方案 | ... | ... |
| 消息中间件 | ... | ... |
| 搜索引擎 | ... | ... |
| API风格 | ... | ... |
| 认证方案 | ... | ... |
| 可观测性 | ... | ... |
| 部署方案 | ... | ... |
### 架构模式 (Architecture Pattern)
[Describe the recommended architecture pattern and why]
### 弹性设计 (Resilience)
[Circuit breaker, retry, bulkhead, etc. specific to this scenario]
### 关键库/依赖 (Key Libraries)
[List with brief explanation of each]
### 潜在风险与权衡 (Risks & Trade-offs)
[Be honest about what you're trading off with this choice]
### 如果规模增长... (Scaling Path)
[How to evolve this architecture as load grows 10x]
After presenting the recommendation, offer to go deeper. Read the relevant reference file based on the recommended language:
references/java-springboot.mdreferences/rust.mdreferences/go.mdThen provide:
If the user's scenario maps to multiple languages (e.g., a polyglot microservice system), present multiple matched rows and explain the trade-offs. For example, a trading platform might use:
Read all relevant reference files and provide guidance for each component.
If the user skips the Q&A and says something like "help me write a REST API in Go" or "set up a Spring Boot microservice", respect that:
references/*.md fileThese apply across all three languages. Language-specific implementation details are in the reference files.
Service Decomposition
Communication Patterns
Resilience
Database Selection
Connection Pooling
pool_size ≈ (core_count * 2) + effective_spindle_countCaching Strategy
Common Bottlenecks (check in this order)
Concurrency Models
429 with Retry-After/health/live + /health/ready| Criteria | Java + Spring Boot 3 | Rust | Go |
|---|---|---|---|
| Best for | Complex business logic, enterprise | Ultra-low-latency, safety-critical | Infra, cloud-native, network services |
| Dev speed | Fast (rich frameworks) | Slower (borrow checker) | Fast (simple language) |
| Memory | Higher (JVM) | Lowest | Low |
| Latency | Good (Virtual Threads) | Best | Very good |
| Ecosystem | Massive | Growing fast | Strong for infra/cloud |
npx claudepluginhub gchensi-2/high-perf-backend-skill --plugin high-perf-backendGuides backend tasks like REST/GraphQL API design, database query optimization, authentication, microservices building, code review, migrations, and load testing across detected stacks.
Designs scalable backend APIs and microservices architectures focusing on service boundaries, data contracts, resilience, observability, and distributed systems. Use for new services or integration planning.
Guides system architecture design including patterns like monolith/microservices, database selection favoring Postgres/Redis, tech stack choices, TypeScript/Next.js project structure, API design, and caching strategies for scale.