context-degradation

Featured

This skill should be used when the user asks to "diagnose context problems", "fix lost-in-middle issues", "debug agent failures", "understand context poisoning", or mentions context degradation, attention patterns, context clash, context confusion, or agent performance degradation. Provides patterns for recognizing and mitigating context failures.

AI & Automation 1,004 stars 164 forks Updated yesterday MIT

Install

View on GitHub

Quality Score: 93/100

Stars 20%
100
Recency 20%
100
Frontmatter 20%
70
Documentation 15%
100
Issue Health 10%
50
License 10%
100
Description 5%
100

Skill Content

# Context Degradation Patterns Diagnose and fix context failures before they cascade. Context degradation is not binary — it is a continuum that manifests through five distinct, predictable patterns: lost-in-middle, poisoning, distraction, confusion, and clash. Each pattern has specific detection signals and mitigation strategies. Treat degradation as an engineering problem with measurable thresholds, not an unpredictable failure mode. ## When to Activate Activate this skill when: - Agent performance degrades unexpectedly during long conversations - Debugging cases where agents produce incorrect or irrelevant outputs - Designing systems that must handle large contexts reliably - Evaluating context engineering choices for production systems - Investigating "lost in middle" phenomena in agent outputs - Analyzing context-related failures in agent behavior ## Core Concepts Structure context placement around the attention U-curve: beginning and end positions receive reliable attention, while middle positions suffer 10-40% reduced recall accuracy (Liu et al., 2023). This is not a model bug but a consequence of attention mechanics — the first token (often BOS) acts as an "attention sink" that absorbs disproportionate attention budget, leaving middle tokens under-attended as context grows. Treat context poisoning as a circuit breaker problem. Once a hallucination, tool error, or incorrect retrieved fact enters context, it compounds through repeated self-reference. A poisoned go...

Details

Author
guanyang
Repository
guanyang/antigravity-skills
Created
10 months ago
Last Updated
yesterday
Language
TypeScript
License
MIT

Integrates with

Similar Skills

Semantically similar based on skill content — not just same category