hadimiftahulf/antigravity-skills

optimize-perf

Systematic performance tuning workflow to identify bottlenecks in Database, Backend, and Frontend layers.

查看源码

使用方法

  1. 复制安装命令,在自己的智能体环境安装前先检查源码包。
  2. 从原始文档核对所需工具、凭据和支持的宿主。
  3. 说明输入材料与期望交付物,并对照来源检查输出。
npx skills add https://github.com/hadimiftahulf/antigravity-skills --skill optimize-perf

开始前需了解

  • 安装不包含模型订阅、API 额度或访问文件的权限。
  • 仓库星数和安装量反映活跃情况,不等于独立质量或安全测试。
仓库原始内容

按源仓库内容呈现,保留标题、案例、代码、表格、链接以及原文引用的演示图片。

Optimize Performance (The Racer 🏎️)

Merged: performance-profiler + optimize-perf.

When to Activate

  • User mentions: "slow", "optimize", "performance", "benchmark", "N+1", "memory".

Phase 1: Profiling (formerly performance-profiler)

Detect anti-patterns before optimizing:

  1. Database: N+1 queries, missing indexes, full table scans, unbounded queries.
  2. Backend: Synchronous IO in request cycle, blocking calls, missing caching.
  3. Frontend: Large bundle size, unoptimized images, excessive re-renders, layout thrashing.
  4. Memory: Object retention, circular references, unbounded lists.

Phase 2: Optimization

After profiling, apply fixes:

  1. Database: Add indexes, eager loading, query optimization, pagination.
  2. Backend: Async processing, queue jobs, Redis/Memcached caching, connection pooling.
  3. Frontend: Code splitting, lazy loading, image optimization, memoization.

Rules

  • Profile FIRST, optimize SECOND. Never guess.
  • Measure before and after. Quantify improvement.
  • One optimization at a time to isolate impact.

Cost: High (Explicit Trigger Required)

按任务相关推荐