debugger_skill

Overview

This skill diagnoses and fixes runtime issues, errors, and bugs through a structured process of reproduction, root-cause analysis, and targeted fixes. I focus on reproducible steps, minimal changes that address the true cause, and validation with tests and builds. It is tuned for TypeScript projects, test suites, performance problems, integration failures, and race conditions.

How this skill works

I start by gathering complete evidence (errors, stack traces, logs, environment). I create a minimal reproduction, form hypotheses about the failure mode, trace execution to the data source, and identify the fundamental cause. Fixes are small, defensive, and tested; validation includes running the original reproduction, full test and build steps, and regression checks.

When to use it

• Runtime crashes, unhandled promise rejections, or exceptions in production
• Failing unit, integration, or E2E tests
• TypeScript compile errors or type mismatches
• Logic bugs producing incorrect outputs or edge-case failures
• Performance problems: slow operations, memory leaks, or excessive renders
• Race conditions, async timing bugs, or state management errors

Best practices

• Create a minimal, repeatable reproduction before changing code
• Collect full error messages, stack traces, and environment details
• Form hypotheses and test them incrementally—avoid random edits
• Make minimal, targeted fixes and add defensive checks where needed
• Add or update tests to prevent regressions and run lint/build/test before merge
• Use structured logging (logger service) instead of leaving console.logs

Example use cases

• A TypeScript compile error caused by a missing null check on API responses
• A flaky test that times out because a mocked function returned a plain object instead of a promise
• A production crash from an unhandled promise rejection in an async useEffect
• Excessive React re-renders due to incorrect useEffect dependencies
• A race condition where state updates occur after component unmount