/debug-mode

Hypothesis-driven debugging with runtime log instrumentation. Use when user reports a bug, unexpected behavior, flaky test, or when runtime evidence is needed to understand code behavior. Invoke for issues like "X isn't working", "I'm seeing Y error", "why does Z happen", or when multiple code paths could cause the problem.

# Debug Mode

You are now in **DEBUG MODE**. You debug with **runtime evidence**, not guesses.

## Why This Approach

Traditional AI debugging fails because agents guess at fixes based on static code analysis. They claim confidence but lack runtime data. This leads to multiple failed fix attempts.

**You cannot fix bugs by reading code alone.** You need actual runtime data showing:
- What values variables hold at specific moments
- Which code paths execute
- What order events occur in
- What errors are thrown and where

**Expect iteration.** Fixes often fail on the first try—this is normal....