debugging_skill

Overview

This skill teaches practical Python debugging using pdb, IDE debuggers, logging, profiling, and error analysis. I guide you from basic breakpoint debugging to production-safe tools like py-spy and structured logging. You will learn to reproduce, inspect, and fix bugs faster and with less disruption to production.

How this skill works

I cover core pdb commands, breakpoints(), post-mortem and remote debugging, plus enhancements like pdb++. I walk through IDE workflows for VS Code and PyCharm: conditional breakpoints, watches, and call stack navigation. I also teach logging setup, profiling with cProfile/line_profiler/memory_profiler, and techniques for analyzing tracebacks and integrating error monitoring.

When to use it

• When you need to step through code to find logic errors or inspect state.
• When investigating production issues without stopping services (py-spy, remote pdb).
• When performance problems require CPU or memory profiling.
• When tracebacks are unclear and you need structured error context.
• When setting up logging and monitoring to reduce future debugging time.

Best practices

• Start with a minimal reproducible example before deep debugging.
• Use conditional breakpoints and watch expressions to limit noise.
• Prefer structured logging and appropriate log levels over print debugging.
• Profile with representative workloads and compare before/after changes.
• Use post-mortem debugging and error monitoring to capture issues in production.

Example use cases

• Step through a failing unit test with pdb to inspect local variables and call flow.
• Attach py-spy to a production process to collect flame graphs without downtime.
• Configure VS Code breakpoints with conditions to skip noisy iterations in a loop.
• Use cProfile and line_profiler to identify a slow function and optimize hotspots.
• Integrate Sentry or a similar service to capture tracebacks and context for uncaught exceptions.