What problem does it solve? Debugging sessions often spiral into guesswork, repeated failed fixes, and lost context when AI conversations reset. This Skill enforces a disciplined investigation process that finds root causes instead of masking symptoms, and preserves debugging state across sessions. ## Core Features & Use Cases - Hypothesis Testing Framework: Forms falsifiable hypotheses, gathers evidence, and eliminates theories systematically while avoiding cognitive biases like anchoring and confirmation bias. - Persistent Debug State: Tracks symptoms, eliminated hypotheses, and evidence in a structured DEBUG.md file so progress survives context resets and session handoffs. - 3-Strike Rule and Restart Protocol: Detects tunnel vision after three failed fix attempts and recommends a fresh context with documented findings. - Use Case: A developer reports that a counter increments twice per click. The debugger forms specific hypotheses (e.g., handleClick fires twice due to double event binding), tests each one, documents eliminated theories, and verifies the fix against a checklist. ## Quick Start Ask the AI to debug the issue where the counter shows 3 after a single click and have it track the investigation in DEBUG.md.
