Runtime Behavior Debugger

Overview

This agent specializes in debugging runtime failures—situations where the application either crashes outright or behaves incorrectly without triggering a compile-time error. It employs a strict 'victim-first' analysis methodology, ensuring that investigation always begins on the device exhibiting the failure.

Capabilities

• Symptom Locking: Forces the capture of precise failure details (mode, summary, affected device) before any technical deep dive.
• Victim-First Analysis: Adheres to the rule of analyzing the failing device first, ignoring sender-side logs initially.
• Failure Window Definition: Systematically identifies the 'last correct state' and the 'first wrong log line' to isolate the exact point of failure in the execution timeline.
• Structured Reporting: Provides a mandatory header format for consistent, actionable debugging reports.

Example Use Cases

• Intermittent UI Glitches: When a feature fails to update the screen correctly (e.g., data doesn't propagate), use this agent to trace the state changes leading up to the visual discrepancy.
• Silent Failures: If a background process stops unexpectedly without an explicit crash, this tool helps define the boundary between normal operation and failure.
• Crash Triage: When a FATAL EXCEPTION occurs, it guides the user through capturing the necessary context (log files, signal) to pinpoint the root cause rather than just reporting the stack trace.