Context Degradation Diagnostics

概要

The Context Degradation skill provides a specialized framework for identifying and resolving the inherent performance drops that occur as LLM context length increases. It equips developers with diagnostic patterns for recognizing the 'lost-in-middle' phenomenon, context poisoning, distraction, and architectural confusion that can derail multi-agent systems. By applying empirical benchmarks and model-specific thresholds for leading LLMs like Claude and GPT, this skill enables the design of resilient context management strategies—such as partitioning, isolation, and recovery—essential for production-grade agentic workflows.

主な機能

• Mitigation strategies for context distraction and irrelevant information noise
• Architectural solutions for context isolation and task segmentation
• Detection of 'lost-in-middle' attention gaps and U-shaped recall curves
• Model-specific performance benchmarks for Claude, GPT, and Gemini
• 7,140 GitHub stars
• Identification and recovery patterns for context poisoning and feedback loops

ユースケース

• Debugging unexpected agent failures in long-running production conversations
• Optimizing RAG system performance by managing information placement and noise
• Designing high-reliability multi-agent architectures with segmented context windows