diagnosing-bugs

What problem does it solve?

This Skill helps you diagnose hard bugs and performance regressions by forcing a tight, red-capable feedback loop before you guess at causes.

Core Features & Use Cases

• Feedback-loop first debugging: Build a failing test, curl check, CLI repro, browser script, trace replay, harness, fuzz loop, or bisect workflow that catches the exact symptom.
• Reproduction and minimization: Confirm the issue reliably, then shrink it to the smallest scenario that still fails so the true cause becomes easier to isolate.
• Hypothesis-driven investigation: Rank falsifiable explanations, instrument one variable at a time, and use targeted logs or debugging tools to separate competing causes.
• Regression protection: Turn the minimized repro into a regression test when the seam is correct, then verify the fix against the original scenario and clean up debug artifacts.
• Use Case: A flaky export failure, a slow API endpoint, or a branch-specific regression can be reduced to a deterministic command that proves the bug before any fix is attempted.

Quick Start

Use the diagnosing-bugs skill to build a tight reproduction loop for the failing behavior, confirm it goes red, and minimize it to the smallest reliable case.