@kiwifruit13

memory-and-context-engineering provides five context engineering primitives (selection, compression, retrieval, state, memory) via a four-layer orchestrator. describes ambitious cognitive modeling with dual-track chains and semantic buckets, though procedural specificity and trigger coverage remain underdeveloped.

agi-evolution-model-basic describes a dual-loop agi framework claiming intentionality analysis, personality mapping, and metacognition. the document is architecturally ambitious but procedurally vague, mixing philosophical abstraction with incomplete technical specs. core workflows reference nonexistent scripts and lack actionable decision points.

基于双环架构的AGI进化模型，通过意向性分析、人格层映射、元认知检测和错误智慧库实现持续自我演进；当用户需要智能对话、人格定制、复杂问题求解或从错误中学习时使用