Agent Skills: Event Loop Skill

Event Loop Skill

Expert skill for high-performance event-driven I/O programming across platforms and runtime environments.

Capabilities

• Platform Multiplexers: Configure epoll (Linux), kqueue (BSD/macOS), IOCP (Windows)
• Performance Analysis: Analyze event loop performance and identify bottlenecks
• Event Debugging: Debug event handling issues and callback delays
• Framework Code Generation: Generate code for libuv, Boost.Asio, Tokio, and mio
• C10K+ Optimization: Optimize for massive concurrent connection handling
• Profiling: Profile event loop performance with system tools
• io_uring Integration: Implement Linux io_uring for high-performance I/O
• Timer Management: Efficient timer wheel and hierarchical timing implementations

Tools and Dependencies

• strace/dtrace - System call tracing
• perf - Linux performance analysis
• libuv - Cross-platform async I/O
• io_uring - Linux async I/O interface
• Tokio - Rust async runtime
• Boost.Asio - C++ async I/O

Target Processes

• event-driven-socket-handler.js
• tcp-socket-server.js
• websocket-server.js
• layer4-load-balancer.js

Usage Examples

epoll Event Loop (C)

int epfd = epoll_create1(0);
struct epoll_event ev, events[MAX_EVENTS];
ev.events = EPOLLIN | EPOLLET;  // Edge-triggered
ev.data.fd = listen_fd;
epoll_ctl(epfd, EPOLL_CTL_ADD, listen_fd, &ev);

while (1) {
    int n = epoll_wait(epfd, events, MAX_EVENTS, -1);
    for (int i = 0; i < n; i++) {
        handle_event(&events[i]);
    }
}

io_uring Setup

struct io_uring ring;
io_uring_queue_init(256, &ring, 0);
// Submit and complete I/O operations

Performance Profiling

perf record -g ./server
perf report --stdio
strace -c -f ./server

Quality Gates

• Event handling latency within threshold
• No callback starvation
• Memory usage stable under load
• CPU utilization optimized
• C10K+ connection handling verified