Performance Analysis

What it does

This skill provides a comprehensive suite of tools for performance analysis within agent workflows. It enables AI agents to identify bottlenecks in communication, processing, memory, and network usage. The skill also facilitates real-time monitoring, historical analysis of swarm operations, automated report generation, and offers AI-powered recommendations for optimizing performance.

When to use it

• When you need to pinpoint the cause of slow or inefficient agent workflows.
• To generate detailed reports on agent performance over specific time periods (e.g., 1 hour, 24 hours, 7 days).
• For automatically identifying and attempting to fix performance bottlenecks within a swarm.
• When you need to analyze the performance of a particular swarm using its ID.

Key capabilities

• Bottleneck Detection: Identifies performance bottlenecks across communication, processing, memory, and network.
• Performance Profiling: Provides real-time monitoring and historical analysis of swarm operations.
• Report Generation: Creates comprehensive performance reports in multiple formats (e.g., HTML).
• Optimization Recommendations: Offers AI-powered suggestions for improving performance.
• Automatic Fixes: Can apply automatic optimizations to detected bottlenecks based on a defined threshold.

Example prompts

• "Run a basic bottleneck detection analysis."
• "Generate a performance report in HTML format, including metrics, covering the last 24 hours."
• "Detect and automatically fix any bottlenecks with a severity above 15% within swarm 'swarm-123'."

Tips & gotchas

• The skill requires claude-flow to be installed.
• You can specify a time range for analysis (e.g., 1h, 24h, 7d, all).
• Adjust the threshold percentage (--threshold) to control the sensitivity of bottleneck detection; lower values will identify more issues but may also include less critical ones.