{"skill":{"slug":"memory-and-context-engineering","displayName":"AGI记忆模组","summary":"用户与模型之间的任何交互行为都可触发；提供Context Engineering五大核心能力（选择、压缩、检索、状态、记忆）及认知模型层支持；作为元技能强制常驻运行","tags":{"latest":"1.0.10"},"stats":{"comments":0,"downloads":190,"installsAllTime":0,"installsCurrent":0,"stars":0,"versions":6},"createdAt":1774494898661,"updatedAt":1776702709214},"latestVersion":{"version":"1.0.10","createdAt":1776701502952,"changelog":"### 1.0.10 · Major structural upgrade & documentation expansion\n\n- Introduced a four-layer architecture with clear separation between control, cognitive, implementation (dual-track), and storage layers; `ContextOrchestrator` as the unified entry point.\n- Refined \"压缩\" logic to cover five chain types (causal, logic, operation, narrative, time); added advanced context compression and result fusion strategies.\n- Added over 30 reference and implementation files, including best practices, troubleshooting, and detailed dual-track/chain logic modules.\n- Enhanced privacy consent process: explicit agent-side consent handling; clarified process in documentation.\n- Improved modularity, expandability, and API robustness; updated interfaces for type safety, error handling, and customization.\n- Expanded and reorganized documentation for architecture, usage examples, API reference, error handling, and integration guides.","license":"MIT-0"},"metadata":{"os":null,"systems":null},"owner":{"handle":"kiwifruit13","userId":"s17cje0hnbtrzjwaa1vrpxyta983ga4v","displayName":"kiwifruit13","image":"https://avatars.githubusercontent.com/u/143737192?v=4"},"moderation":{"isSuspicious":true,"isMalwareBlocked":false,"verdict":"suspicious","reasonCodes":["suspicious.llm_suspicious","suspicious.privileged_always"],"summary":"Detected: suspicious.llm_suspicious, suspicious.privileged_always","engineVersion":"v2.4.0","updatedAt":1776702709214}}