pattern·free·context_pollution
Context pollution
Memory loaded too greedily floods every prompt with marginally-relevant context. AI loses focus, misses the point, hallucinates trying to integrate everything.
Especially common with markdown-based memories that load in full instead of being queried selectively. Symptoms: AI starts answers with "considering all the context, let me address each point". Mitigations: cap CLAUDE.md at 500 lines, use semantic retrieval (structured memory servers do this automatically), tier context (L0 summary / L1 standard / L2 full detail — only load what is relevant to current prompt).
Beziehungen
memoryriskscontext-windowpattern