Openclaw Never Forget Skill (全知记忆管理引擎)

🎯 核心愿景 (Core Vision)

This is a unified memory system designed to transform Openclaw from a stateless assistant into an evolving entity with continuous, reliable context. It merges short-term anti-amnesia tactics (periodic snapshots) with long-term knowledge structuring (semantic extraction and daily logs). 这是一个统一的记忆系统，旨在将 Openclaw 从无状态助手转变为具有连续、可靠上下文的不断进化的实体。它将短期防遗忘战术（定期快照）与长期知识结构化（语义提取和每日日志）结合在一起。

🧠 记忆层级架构 (Memory Architecture Tiers)

1. Episodic Memory (情景记忆) - memory/YYYY-MM-DD.md

   • What: The chronological journal of the day. (当天的按时间顺序记录的日志表。)
   • Contents: Task logs, errors encountered, successful actions, and 30-minute short-term context summaries. (任务执行日志、遇到的错误、成功的操作以及每30分钟的短期上下文总结。)

2. Semantic Memory (语义记忆) - memory/MEMORY.md (or core knowledge files)

   • What: The distilled, timeless facts and rules learned about the user and their projects. (关于用户及其项目的经过提炼的、永久的事实和规则。)
   • Contents: Preferences, recurring workflows, API keys context (not raw), important structural decisions. (偏好设置、重复出现的工作流、重要的结构性决策。)

⚙️ 触发与运行机制 (Triggers & Execution Logic)

阶段 1: 隐性后台运转 (Phase 1: Silent Background Operations)

Trigger: Continuous during active sessions. (活跃对话期间持续进行。)

• 00:00 Daily Boot (每日零点初始化):
  • On the first prompt of a new day, verify the existence of memory/YYYY-MM-DD.md. Create it with a standard template if missing.
  • (在每天的第一次对话时，验证并初始化当天的记忆文档。)
• 30-Minute Checkpoint (30分钟强制快照):
  • Action: Every 30 minutes of active interaction, briefly summarize the immediate context (What are we building? What bugs are we fixing?).
  • Storage: Append this snapshot seamlessly under the current time block in today's Episodic Memory file.
  • (每活跃30分钟，简要总结当前的即时上下文，并静默追加到当天的情景记忆文件中，以此防止长文本窗口溢出。)
• Insight Extraction (洞察提取):
  • Action: If a significant decision is made, a rule is established by the user, or a major task completes, extract the "Core Knowledge" and append it to the long-term memory/MEMORY.md.
  • (如果做出了重大决定、用户确立了规则或重大任务完成，自动提取“核心知识”并追加到长期的 memory/MEMORY.md 中。)

阶段 2: 显性主动检索 (Phase 2: Explicit Active Retrieval)

Trigger: Explicit user commands like "let openclaw recall previous chats", "what did we do yesterday?", "review my preferences", "让openclaw回忆之前的聊天", "我们昨天做了什么". (明确的用户指令触发。)

• Action: Use tools like memory_get or memory_search (if available in your environment) to read the relevant YYYY-MM-DD.md files or the global MEMORY.md.
• Synthesis: Do not just paste raw logs. Analyze the retrieved data and provide a coherent, synthesized summary bridging the past memory with the current prompt's needs.
• (使用文件读取机制快速调取相关日期的记忆文件或全局知识库。分析提取的数据，提供一个连贯、综合的总结，将过去的记忆与当前提示的需求连接起来。)

📝 存储规范与模板 (Storage Standards & Templates)

1. Daily Episodic Template (memory/YYYY-MM-DD.md)

# Openclaw Daily Log - [YYYY-MM-DD]