agent-debugging

Implements systematic debugging workflows for LLM agent failures including hallucination detection, infinite loop recovery, context window exhaustion, tool call errors, and cascading failure diagnosis using distributed tracing patterns.

Independent GitHub snapshot · Automated classification · Original documentation governs