Optimiser les performances

Ce que contient l'archive

1 outil, prêt à déposer dans votre projet :

• performance-profiler — Use this agent when you need to diagnose and fix performance bottlenecks in your application. Specifically invoke this agent when: (1) API endpoints or views exceed their latency budgets (e.g., p95 > 500ms), (2) CPU or memory spikes are detected in monitoring dashboards, (3) specific code paths are suspected of causing slowdowns, or (4) you need to optimize critical user-facing features before launch.\n\nExamples:\n\n\nContext: Developer notices slow API response times in production monitoring.\nuser: "Our /api/search endpoint is averaging 2 seconds response time, way above our 500ms budget"\nassistant: "I'm going to use the Task tool to launch the performance-profiler agent to analyze this endpoint and identify the bottlenecks."\nThe user has identified a specific performance issue exceeding latency budgets, which is a clear trigger for the performance-profiler agent.\n\n\n\nContext: After implementing a new feature, the developer wants proactive performance validation.\nuser: "I just added batch processing for user notifications. Here's the code:"\n\nassistant: "Let me use the performance-profiler agent to analyze the performance characteristics of this new batch processing code and ensure it won't create bottlenecks."\nProactive performance profiling after significant code changes helps catch issues before they reach production.\n\n\n\nContext: Monitoring alert shows memory spike.\nuser: "Getting alerts about memory usage jumping to 85% during peak hours"\nassistant: "I'll use the Task tool to launch the performance-profiler agent to profile memory allocation patterns and identify what's causing these spikes."\nMemory pressure is a performance issue that requires profiling to diagnose root causes.\n