🔧 Debugging

Systematic bug investigation with hypothesis tracking

MOOLLM K-Lines

K-Line	Why Related
adventure/	Debugging IS adventure + hypothesis tracking
sniffable-python/	Readable code aids debugging
scratchpad/	Debug scratch notes
research-notebook/	Investigation notes
sister-script/	Document the fix
session-log/	Track all debug steps
self-repair/	Automated fixes
play-learn-lift/	Debugging IS learning
constructionism/	Try → Inspect → Hypothesize → Modify

Quick Links:

• Full Specification — complete protocol and schemas

Overview

Debug problems methodically. Track hypotheses, test systematically, converge on root causes. Treats debugging as a structured adventure where each hypothesis is a quest to validate or refute.

The debugging loop: OBSERVE → HYPOTHESIZE → TEST → LEARN → FIX

Key Techniques

Technique	When to Use
Binary Search	Bug is somewhere in large space
Rubber Duck	Stuck, need fresh perspective
Minimal Repro	Complex system, unclear cause
Git Bisect	Bug is a regression
Print Debug	Need to understand flow

Example Session

debug:
  bug: "Login fails with valid credentials"
  
  hypotheses:
    - id: 1
      claim: "Password hashing mismatch"
      status: refuted
      evidence: "Hashes match in logs"
      
    - id: 2
      claim: "Session cookie not set"
      status: confirmed
      evidence: "Cookie missing in response headers"
      
  fix: "Add Set-Cookie header to auth response"
  verified: true