ZHI-QI/dsh-okf-memory

README 摘要

dsh-okf-memory 简体中文 English Session-to-OKF memory plugin with neuro-self-learning: predictive recall, uncertainty-driven capture, reinforcement feedback, consolidation & forgetting. Turn high-value content from your conversations into persistent long-term memory, organized as OKF v0.1 knowledge documents. The agent gets smarter the more you use it — every selection, skip, and correction is a learning signal that updates memory weights. Features - Four-stage memory loop : Capture → Concept-ize (OKF) → Consolidate → Recall - OKF v0.1 compliant : every concept is a standard Markdown document (frontmatter hard-requires type ), index.md progressive catalog + log.md change history, cross-links use bundle-absolute paths - Neuro-self-learning driver : predictive recall (predict first, then verify by retrieval), uncertainty-driven exploration (expand search when confidence is low), prediction-error-driven capture (user corrections / first-time disclosures / counter-intuitive conclusions trigger writes), weight decay + archiving (consolidation & forgetting) - Reinforcement feedback loop : score = relevance × weight × recency ; selecting a candidate raises its weight, skipping lowers it - Tec…

分类