debug-systematiclisted

# Systematic Debugging Workflow I'll help you debug issues systematically using the scientific method - hypothesis formation, testing, and iterative refinement. Arguments: `$ARGUMENTS` - error description, reproduction steps, or context ## Token Optimization **Target:** 50% reduction (4,000-6,000 → 1,500-3,000 tokens) ### Core Optimization Strategies **1. Hypothesis-Driven Debugging (Not Exhaustive Analysis)** - ❌ **AVOID:** Reading entire codebase to find bugs - ✅ **DO:** Form hypotheses about likely causes, test top 2-3 first - **Token savings:** 90% (200 tokens vs 2,000+ tokens) - **Pattern:** Prioritize recently changed files, common failure patterns **2. Git Diff for Recently Changed Files (Likely Bug Source)** - ❌ **AVOID:** `ls -R` then reading all files - ✅ **DO:** `git diff --name-only HEAD~3..HEAD` to find changed files - ✅ **DO:** `git log --oneline --since="3 days ago"` for recent commits - **Token savings:** 85% (300 tokens vs 2,000+ tokens) - **Pattern:** Bugs often introduced in recent changes **3. Stack Trace Parsing with Grep** - ❌ **AVOID:** Reading entire log files with Read tool - ✅ **DO:** `grep -i "error\|exception\|fatal" logs/*.log | tail -20` - ✅ **DO:** Parse stack traces to extract file paths and line numbers - **Token savings:** 95% (100 tokens vs 2,000+ tokens for large logs) - **Pattern:** Stack traces reveal exact failure locations **4. Test Failure Analysis Caching** - ✅ Cache test results in `debug/state.json` - ✅ Cache hypothesis out