← Back to Skills Marketplace
luis1213899

Dream Selfimproving

by luis1213899 · GitHub ↗ · v4.2.1 · MIT-0
cross-platform ⚠ suspicious
318
Downloads
1
Stars
1
Active Installs
33
Versions
Install in OpenClaw
/install dream-selfimproving
Description
让AI拥有进化能力——每晚自动复盘当天对话,提取洞察,更新记忆,像梦境一样在夜里悄然进化。不用手动记录,AI自己记住重要的事,越用越聪明。
README (SKILL.md)

Dream Self-improving — 夜间记忆蒸馏与自我进化

🧠 v4.x 已集成 Long-Term RAG — MetaGPT风格短长记忆合并,条目老化后自动晋升到 RAG 层

现状

Phase 3(✅ 已实现): OpenClaw Hook hippocampus 监听每条消息,实时写入 memory/logs/ Phase 3(✅ 已实现): dream.py v4.x 定时蒸馏 + M-FLOW Bundle Search检索 Phase 3(✅ 已实现): dream.py v4.x Long-Term RAG 长记忆层


核心升级:Long-Term RAG

参考 MetaGPT 的 RoleZeroLongTermMemory 设计,新增短长记忆合并机制:

short-term-recall.json  ←  活跃recall条目(上限200条)
memory/.rag/longterm.jsonl  ←  老旧条目RAG存储

晋升条件:

  • 条目 age > 30天(从最后召回时间算)
  • 且 recallCount \x3C 3(未被频繁召回)

召回流程:

  1. 蒸馏前,从当日高权重条目提取关键词
  2. 用关键词查询 RAG,召回相关旧记忆
  3. 旧记忆注入蒸馏上下文,让 AI 知道"之前有过什么"

效果: 记忆越来越精准,不像以前每次都从零开始。


完整链路

用户对话
   ↓
OpenClaw Hook: message:preprocessed
   ↓
丘脑过滤(Thalamus)→ 杏仁核标记(Amygdala)→ 海马体存储(memory/logs/)
   ↓
cron 触发(早7点/晚10点)
dream.py v4.x
   ↓
[4.5] RAG查询 — 从当日条目提取关键词 → 查询memory/.rag/longterm.jsonl → 注入蒸馏上下文
   ↓
Bundle Search检索(替代简单grep)
   ↓
杏仁核标记融合 → Auditor审计 → 分析皮层模式识别 → 前额叶蒸馏规划
   ↓
[4.6] RAG晋升 — 30天+未召回条目 → 写入longterm.jsonl
   ↓
归档区 → 真相文件写回 → 梦境报告

M-FLOW 核心架构

倒锥知识图谱(Inverted Cone)

所有记忆组织为四层有向图,形成倒锥结构:

          锥尖(容易精确命中)
             ↓
    ┌─────────────────────────┐
    │  L4 Entity             │  ← 用户/项目/系统等实体节点
    │  L3 FacetPoint         │  ← 具体属性、特征、标签
    │  L2 Facet               │  ← 一组相关特征
    │  L1 Episode(锥底)     │  ← 最终返回的知识单元
    └─────────────────────────┘
          锥底(返回给用户)

搜索逻辑(Bundle Search):

  1. 锥尖广撒网:查询向量化后同时在4层搜索,每个集合返回最多100个候选
  2. 投影到图中:命中点作为入口,提取周围子图(边+邻居+连接关系)
  3. 代价传播:沿边从锥尖向锥底传播,Episode得分 = 所有路径中最小代价

三条核心设计原则:

原则 说明 对应效果
边携带语义 每条边附带自然语言描述,参与检索 不是被动连接,是主动语义过滤器
路径最小代价 一条强证据链就足以证明相关性 不被无关路径稀释分数
惩罚直接命中Episode 直接匹配摘要反而加惩罚 偏好精准锚点路径,防止宽泛匹配

脑区协同架构

① 丘脑(Thalamus)— 注意力门控

过滤纯问候/简单确认,只记录有意义的事件 标记类型:event / decision / correction / completed / insight / error

② 杏仁核(Amygdala)— 情绪标记

correction/error/decision/completed/insight 携带 HIGH 权重,优先蒸馏

③ 海马体(Hippocampus)— M-FLOW图存储 + RAG

Phase 1:memory/logs/ 追加日志(Episode层)

Phase 2:构建M-FLOW图结构:

Episode (L1)              ← daily log / topic file
  ↓ semantic edge
Facet (L2)                ← grouping: correction_group, project_xxx
  ↓ semantic edge  
FacetPoint (L3)           ← specific tag: error.timeout, user.pref
  ↓ semantic edge
Entity (L4)               ← user, project, tool, skill

FacetPoint = type + topic + keywords 的向量描述(向量化后参与Bundle Search) 语义边描述 = "这个FacetPoint为什么属于这个Episode" 的自然语言说明

④ 前额叶(Prefrontal Cortex)— Bundle Search + RAG召回 + 蒸馏规划

Bundle Search检索替代简单grep:

查询 → 向量化 → 4层锥形搜索 → 代价传播 → 最小路径Episode

RAG召回(v4.x新增):

当日关键词 → 查询longterm.jsonl → 召回相关旧记忆 → 注入蒸馏上下文

⑤ 蓝斑核(Locus Coeruleus)— 警觉与新鲜度信号

freshness分数——最近被提及的记忆权重更高


Long-Term RAG Layer 详解

存储结构

memory/
├── .dreams/
│   └── short-term-recall.json   # 活跃recall条目(上限200条)
└── .rag/
    └── longterm.jsonl          # 老旧条目RAG存储(JSONL格式)

晋升机制

# 晋升条件
if age_days > 30 and recall_count \x3C 3:
    promote_to_longterm_rag(entry)

召回机制

# 蒸馏前
keywords = [v['snippet'][:100] for v in tagged.values()][:20]
query = ' '.join(keywords[:5])
rag_results = query_longterm_rag(query, k=5)

# 召回结果注入蒸馏上下文
learnings['LEARNINGS.md'] += f"\
\
## Long-Term Memory (RAG)\
{rag_text}"

手动命令

# 查看短/长记忆状态
python skills/dream-selfimproving/scripts/longterm_rag.py --status

# 手动晋升老条目
python skills/dream-selfimproving/scripts/longterm_rag.py --promote

# 搜索长记忆
python skills/dream-selfimproving/scripts/longterm_rag.py --query "关键词"

Pattern Library

Patterns are reusable response templates extracted from recurring learnings:

memory/patterns/
└── p-xxx.md           # Pattern files with trigger + response

Pattern格式(含M-FLOW元数据):

---
name: pattern名称
trigger: 什么情况下触发
response: 如何响应
examples: [案例1, 案例2]
created: YYYY-MM-DD
updated: YYYY-MM-DD
# M-FLOW 元数据
entity: pattern          # L4 Entity
facets: [tag1, tag2]    # L3 FacetPoints
episode_id: p-xxx        # L1 Episode
---

Memory Taxonomy & M-FLOW映射

Memory Type L4 Entity L3 FacetPoints L1 Episode
user user.luyi role, pref, goal, communication_style topics/user_*.md
feedback feedback correction, error, insight, confirmation topics/feedback_*.md
project project.{name} decision, tool, deadline, context topics/project_*.md
reference reference credential, link, skill, system topics/reference_*.md
longterm (RAG) aged, promoted .rag/longterm.jsonl

Directory Structure (v4.x)

memory/
├── graph/                         # M-FLOW 知识图谱
│   ├── entities.json              # L4 Entity 节点列表
│   ├── facetpoints.json           # L3 FacetPoint 节点列表
│   ├── facets.json                # L2 Facet 节点列表
│   ├── episodes.json              # L1 Episode 节点列表
│   ├── edges.json                 # 语义边(含描述文本)
│   └── index.json                 # 图索引 + 向量锚点
├── logs/
│   └── YYYY/MM/YYYY-MM-DD.md     # Daily append-only logs (Episode)
├── topics/                        # Distilled topic memories
│   ├── user_xxx.md
│   ├── feedback_xxx.md
│   ├── project_xxx.md
│   └── reference_xxx.md
├── patterns/                      # Pattern Library
│   └── p-xxx.md
├── episodes/                      # Project narratives
├── .dreams/
│   └── short-term-recall.json     # 活跃recall条目(上限200条)
├── .rag/
│   └── longterm.jsonl             # Long-Term RAG(v4.x新增)
├── procedures.md                  # Workflow preferences
├── archive.md                     # Compressed old entries
├── dream-log.md                   # Dream cycle reports
└── MEMORY.md                      # INDEX only

.learnings/                        # self-improving-agent
├── LEARNINGS.md
├── ERRORS.md
└── FEATURE_REQUESTS.md

Health Score (v4.x)

Metric Weight Formula
Freshness 0.20 entries_referenced_last_30_days / total
Coverage 0.20 categories_updated_last_14_days / 10
Coherence 0.20 entries_with_semantic_edges / total
Graph Connectivity 0.20 connected_components_ratio
Efficiency 0.10 max(0, 1 - line_count/500)
Reachability 0.10 Bundle Search路径覆盖率

Dream Distillation Steps (v4.x)

When cron triggers:

  1. Bundle Search预热:用今日日志构建临时图结构,快速验证图连通性
  2. Read memory/logs/{date}.md
  3. Read .learnings/LEARNINGS.md, .learnings/ERRORS.md, .learnings/FEATURE_REQUESTS.md
  4. Read MEMORY.md, topic files, graph/index.json, procedures.md for context
  5. Snapshot BEFORE: count entries, decisions, lessons, procedures
  6. [4.5] RAG召回:从当日条目提取关键词 → 查询longterm.jsonl → 注入蒸馏上下文
  7. 图增强检索:对每个learnings entry执行Bundle Search,找到相关Episode
  8. Distillation Agent: Run sub-agent on raw entries + learnings + RAG results → produce:
    • 3-5 genuine insights ("I learned that...")
    • 1-3 tomorrow action items
    • 0-3 topic files to write to memory/topics/
    • Health metric interpretation
  9. [4.6] RAG晋升:30天+未召回条目 → 写入longterm.jsonl
  10. 更新图结构
    • 新Episode写入 graph/episodes.json
    • 新FacetPoint写入 graph/facetpoints.json
    • 新边写入 graph/edges.json(含语义描述)
  11. Write topic files (from Distillation Agent output)
  12. Update truth files (user_state.md, pending.md)
  13. Update graph/index.json entry metadata + 重新计算向量锚点
  14. Compute health metrics → update graph/index.json stats
  15. Archive eligible entries → append to archive.md
  16. Update MEMORY.md index (max 200 lines)
  17. Snapshot AFTER: calculate deltas
  18. Write dream report to memory/dreams/{date}.md and dream-log.md
  19. [Optional SwarmRecall]: 如果配置了API key,执行云端图同步

Dream Report Format (v4.x)

# 🌙 Dream Report — {date}

## M-FLOW Graph Status
- Entities: N | FacetPoints: N | Episodes: N | Edges: N
- Graph Connectivity: {score}% | Avg Path Cost: {cost:.3f}

## RAG Status
- Short-term recall: N 条 | Long-term: M 条
- Promoted this cycle: N 条

## Health Insights
- {insight based on graph connectivity / Bundle Search coverage}

## Insights ("I Learned")
- {genuine insight 1}
- {genuine insight 2}

## Tomorrow's Focus
- {actionable item 1}
- {actionable item 2}

## Topic Files Written
- {filename}: {title}

## Graph Updates
- New episodes: N
- New semantic edges: N
- Pruned nodes: N

## Analysis
- Recurring errors found: {list}
- Root causes identified: {analysis}
- Bundle Search paths evaluated: {count}

## Patterns Updated
- {pattern_name}: {change}

User Prompts

  • "dream report" / "梦境报告" → read and display latest dream report
  • "dream" / "做梦" → run distillation now
  • "/dream status" → show M-FLOW graph stats, health score, pattern count
  • "/dream search {query}" → run Bundle Search and show top results
  • "/dream rag status" → show RAG status (from longterm_rag.py)


🔒 安全说明

update-cron-date.py 的 --confirm 标志

此脚本用于修改 OpenClaw cron 作业,属于特权操作

  • --confirm 标志是安全机制,防止意外执行
  • 无此标志时,脚本只显示将要做什么但不会实际修改
  • 使用前请确保您了解 cron 作业的当前状态:
    openclaw cron list
    

SWARMRECALL_API_KEY 环境变量

  • 声明位置: metadata.openclaw.requires.env.SWARMRECALL_API_KEY
  • 用途: 可选。云端去重/矛盾检测(需连接 SwarmRecall 服务)
  • 默认值: 不设置则仅本地运行,无云端功能
  • 获取方式: 向 SwarmRecall 服务注册获取

文件访问范围

此技能会访问和修改以下位置的文件:

  • ~/.openclaw/workspace/memory/logs/ — 对话日志
  • ~/.openclaw/workspace/memory/.dreams/ — 梦境报告
  • ~/.openclaw/workspace/memory/.rag/ — 长记忆 RAG 存储
  • ~/.openclaw/workspace/memory/.truth/ — 真相文件
  • ~/.openclaw/workspace/memory/.learnings/ — 学习记录
  • ~/.openclaw/hooks/hippocampus/ — Hook 配置
  • OpenClaw cron 条目(通过 update-cron-date.py)

Scripts

  • dream.py — Phase 2 蒸馏脚本(v4.x,M-FLOW Bundle Search + RAG召回/晋升)
  • update-cron-date.py — 每日 cron 日期注入 ⚠️ 特权操作,需 --confirm
  • graph-builder.py — 从日志构建M-FLOW图结构
  • bundle-search.py — Bundle Search检索实现
  • longterm_rag.py — Long-Term RAG 管理脚本(v4.x新增)

Phase 1 启用(hippocampus hook)

Hook 目录: ~/.openclaw/hooks/hippocampus/ 已配置: openclaw.jsonhooks.internal.entries.hippocampus: enabled: true

功能: 监听 message:preprocessed 事件,自动记录对话到 memory/logs/YYYY/MM/YYYY-MM-DD.md

丘脑过滤规则:

  • 纯问候 / 简单确认(\x3C20字)不记录
  • 高权重标记:correction / error / decision / completed / insight

重启 gateway 后生效:

schtasks /run /tn "OpenClaw Gateway"

M-FLOW vs 旧架构对比

维度 旧架构(平坦检索) M-FLOW(倒锥图路由)
存储结构 平面文件列表 四层有向图
检索方式 grep / 向量相似度 Bundle Search代价传播
关系表示 简单link引用 带语义描述的边
短长记忆 无分层 30天老化晋升RAG

与MetaGPT对比

维度 MetaGPT RoleZeroLongTermMemory Dream Long-Term RAG
RAG引擎 Chroma + LLMRanker JSONL + 关键词匹配
召回触发 memory_k 溢出 或 用户需求 每次蒸馏前
晋升条件 count > memory_k age > 30天 且 recallCount \x3C 3
向量化 embedding 模型 词袋模型(简化版)
复杂度 依赖 Chroma/llama-index 纯 Python,无外部依赖
Usage Guidance
简明建议: 1) 如果你只想试用记忆/蒸馏功能,先在隔离的 workspace 里运行(设置 OPENCLAW_WORKSPACE 指向一个空目录),不要用默认主目录下的 .openclaw/workspace;备份现有 memory/ 数据。 2) 在允许技能运行前,手动审阅下列关键文件(示例): scripts/update-cron-date.py(或任何包含 openclaw cron edit 的脚本)、scripts/dream.py(主蒸馏流程)、graph-builder.py(图生成与写入逻辑)、bundle-search.py、scripts/longterm_rag.py。确认没有你不期望的网络/外部提交或敏感环境读取。 3) 禁用或谨慎对待调度修改:技能提供修改 cron 的能力(需要 --confirm),不要赋予该脚本自动执行 --confirm 的权限;若可能,禁用自动修改 cron,改为手动设置系统/agent 的定时任务。 4) 环境变量与凭证:不要在未审查前向技能提供任何云或第三方凭证(例如 SWARMRECALL_API_KEY)——仅在确知用途并信任代码后才填写。注意 SKILL.md 提到会通过 openclaw env get 动态读取环境,审慎授权。 5) 控制自主运行:如果平台允许,关闭或限制该技能的自动/无人监督调用(要求用户确认每次运行),以避免模型在无人监督时修改调度或写入长期记忆。 6) 运行监控与最小权限:首次运行时在受控环境中执行并监控文件修改、子进程调用与任何外部网络请求(网络请求在代码中未显式看到,但仍建议监控)。如非必要,可移除/注释掉修改 cron 的代码段与任何对 openclaw env get 的调用。 若你希望我进一步检查具体文件(例如完整查看 scripts/update-cron-date.py 或 scripts/longterm_rag.py),把这些文件内容粘贴过来我可以做更精确的风险分析。
Capability Analysis
Type: OpenClaw Skill Name: dream-selfimproving Version: 4.2.1 The skill bundle implements a complex 'self-evolving' memory system (M-FLOW) that distills daily logs into long-term knowledge. It possesses high-privilege capabilities, specifically the ability to modify the agent's cron jobs via 'update-cron-date.py' and append behavioral/prevention rules to 'TOOLS.md' via 'dream.py'. While these actions are aligned with the stated purpose of autonomous self-improvement and include safety mechanisms like the '--confirm' flag, the capacity to alter system-level configurations and the agent's tool definitions represents a significant risk profile. The presence of several self-patching scripts (e.g., 'fix_bundle.py', 'fix_insert.py') that modify the skill's own Python code further increases the complexity and risk, although no explicit evidence of malicious intent or data exfiltration was found.
Capability Tags
requires-sensitive-credentials
Capability Assessment
Purpose & Capability
名称、描述与代码/文档中实现的功能基本一致:收集对话日志、构建 M-FLOW 知识图谱、做短/长记忆合并、运行蒸馏与 RAG、以及提供查询/晋升/归档命令。所含脚本(dream.py、graph-builder.py、bundle-search.py、longterm_rag.py 等)与记忆管理、检索和蒸馏核心功能直接相关,没有发现与描述无关的恰当性缺口。
Instruction Scope
SKILL.md 与脚本会直接读写用户 workspace 下的 memory/*、.learnings、.rag/longterm.jsonl 等文件,这是实现记忆功能所需但范围较大;更重要的是说明与脚本会通过 openclaw CLI(openclaw env get / openclaw cron edit)动态获取环境与编辑 cron。修改 cron(通过 scripts/update-cron-date.py)为特权操作,且 openclaw env get 表示运行时可能读取未声明的环境变量/凭据。SKILL.md 虽有提醒需使用 --confirm,但指令级别仍允许修改系统调度和访问环境信息,授予较高权限。
Install Mechanism
无外部下载或安装规范(instruction-only + 自带 Python 脚本),没有从不可信 URL 拉取可执行文件。风险较低于远程提取,但包含多份本地 Python 脚本会在运行时被执行——这比纯文档技能更有攻击面,建议手动审阅脚本内容并在沙箱/独立 workspace 中运行。
Credentials
声明的环境变量只有可选 SWARMRECALL_API_KEY(用于云端去重/矛盾检测),这与功能相符。问题在于 SKILL.md 明确会通过 `openclaw env get` 动态读取运行时环境,这可能让技能在执行时访问未在元数据中声明的其它环境变量/凭证(例如 AGENT/WOKRSPACE 相关变量或系统级凭证),扩大了凭证暴露面。脚本也默认读取 OPENCLAW_WORKSPACE 或用户主目录下的 .openclaw/workspace,需保证路径隔离。
Persistence & Privilege
技能不会被强制 always:true,但允许模型自主调用(平台默认)。当技能被允许自主运行时,它包含能修改 cron 的脚本(update-cron-date.py / openclaw cron edit),并会写入 workspace 下的长期记忆文件。模型在无人监督下执行这些写入与调度修改会扩大滥用或误操作的影响 —— 特别是若攻击者诱导模型触发带有 --confirm 的 cron 编辑命令。建议限制自动化权限或审查并手动批准任何调度修改。
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install dream-selfimproving
  3. After installation, invoke the skill by name or use /dream-selfimproving
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v4.2.1
安全更新:(1) 明确声明 SWARMRECALL_API_KEY 环境变量 (required:false) (2) 添加安全说明章节,解释 --confirm 标志原因和文件访问范围
v4.2.0
修复硬编码路径问题:WORKSPACE 默认值从 'C:/Users/26240/.openclaw/workspace' 改为跨平台路径 Path.home() / '.openclaw' / 'workspace'
v4.1.6
- 优化描述和标签,使技能定位更明确、内容更精炼。 - 在 metadata.requires 下新增 SWARMRECALL_API_KEY 环境变量说明,支持可选云端去重/矛盾检测。 - “AI记忆进化助手”重命名为“dream-AI进化的梦之核”,副标题和标签更符合功能定位。 - 删除冗余内容,收紧前言,强调“夜晚进化、自动提取记忆”核心卖点。 - 使 cron 环境变量说明与脚本一致,并简化运行说明。
v4.1.5
No file or documentation changes detected for version 4.1.5. - No updates or modifications have been made in this release. - Behavior, functionality, and documentation remain unchanged from the previous version.
v4.1.4
No changes detected in this release. - No updates or modifications to files. - Functionality and documentation remain unchanged from the previous version.
v4.1.3
- Description and metadata simplified for broader audience; technical jargon and complex terms removed from the description. - Added meta fields: human-friendly name, tagline, category, and tags for easier discovery. - Core skill details and architecture remain unchanged; technical documentation preserved for advanced users. - No functional/code changes—documentation readability and accessibility improved.
v4.1.2
No changes detected in this version. - No file or documentation changes since the previous version.
v4.1.1
**v4.1.1 introduces a Long-Term RAG (Retrieval-Augmented Generation) layer for smarter memory management and recall.** - Added Long-Term RAG: Old, infrequently recalled memories are promoted from short-term to a dedicated long-term RAG storage (.rag/longterm.jsonl). - Bundle Search and dream.py now perform keyword-based queries on long-term memory, injecting relevant old knowledge into nightly distillation for improved context awareness. - Promotion and recall logic is automated but also manually controllable via the new scripts/longterm_rag.py utility (supports --status, --promote, --query). - Updated documentation to reflect the new long-term memory flow, storage structure, and integration with the existing M-FLOW knowledge graph system. - Existing directory and health scoring structures expanded to track and score long-term memory usage and coverage.
v4.1.0
修复SKILL.md元数据:name/描述中文化,移除无效env声明消除缺依赖警告;bundle-search单字查询bug修复(无L3锚点时L1直接命中无法传播问题)
v4.0.5
Merge dream into dream-selfimproving; remove 8 obsolete/stale patch files; full syntax validation passed
v4.0.4
dream-selfimproving v4.0.4 - New: After each Dream run, distilled insights and health scores are synced to DREAMS.md for use in the native Dreaming UI. - SKILL.md: Updated phase descriptions and process flow to clarify DREAMS.md integration and Phase 3. - Minor metadata and description enhancements for improved clarity. - No changes to core M-FLOW architecture or distillation logic.
v4.0.3
Version 4.0.3 - Added new file: check_skill.py - Introduces a script or tool for skill checking or validation - No other feature or logic changes documented
v4.0.2
**dream-selfimproving v4.0.2 — Minor improvements and code refactor** - Updated graph-builder.py and scripts/dream.py for better maintainability. - Improved internal logic related to memory graph building and retrieval processes. - Refined interaction between memory distillation (dream.py) and graph storage. - Bug fixes and code clean-up for increased stability.
v4.0.1
dream-selfimproving v4.0.1 - Added _meta.json metadata file for enhanced skill management. - Updated scripts/dream.py (details not visible here) to support latest compatibility or internal logic. - No user-facing feature changes. - Overall structure and docs (SKILL.md) remain unchanged from v4.0.
v4.0.0
v4.0: M-FLOW inverted-cone graph + Recall Store enrichment + Bundle Search + 5 brain-region architecture
v3.3.0
正确的 slug 名称
v3.2.0
dream-selfimproving v3.2.0 - Upgraded nightly memory distillation (`dream.py v3.2`) with full integration of recall store and hippocampus logs for unified processing. - Improved architecture and documentation for multi-agent and multi-file (recall store) memory distillation. - Added detailed explanation of brain-region collaboration and end-to-end flow, including new diagrams and audit steps. - Introduced references/multi-agent-team-architecture.md for extended and team scenarios. - Refined event weighting (Amygdala) and pattern extraction criteria for more precise topic and pattern promotion. - Miscellaneous documentation enhancements, clarifying directory structure and workflow phases.
v3.1.0
dream-selfimproving v3.1.0 - Updated internal Python script (scripts/dream.py) for improved memory processing. - No user-facing changes documented in the changelog. - Retains the same skill architecture and documented interface.
v3.0.0
### Dream Self-Improving Skill v3.0.0 - Introduced `package.json` for modern dependency and metadata management. - Major code revision in `scripts/dream.py` for improved memory architecture. - Refined documentation to clarify the 5-region brain-inspired architecture. - Enhanced health and importance scoring logic and documentation. - Improved integration notes for environment variables and SwarmRecall options.
v2.9.0
v2.9.0: Scientific brain region mapping — expanded from 4 to 5 brain regions based on real neuroscience: (1) Thalamus = attention gating, (2) Amygdala = emotional tagging (corrections/high-weight), (3) Hippocampus = raw log storage, (4) Prefrontal Cortex = distillation planning + circuit breakers, (5) Locus Coeruleus = freshness signals for health scoring. Phase 1: all three work together in real-time. Phase 2: prefrontal coordinates analyst + archiver.
Metadata
Slug dream-selfimproving
Version 4.2.1
License MIT-0
All-time Installs 1
Active Installs 1
Total Versions 33
Frequently Asked Questions

What is Dream Selfimproving?

让AI拥有进化能力——每晚自动复盘当天对话,提取洞察,更新记忆,像梦境一样在夜里悄然进化。不用手动记录,AI自己记住重要的事,越用越聪明。 It is an AI Agent Skill for Claude Code / OpenClaw, with 318 downloads so far.

How do I install Dream Selfimproving?

Run "/install dream-selfimproving" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is Dream Selfimproving free?

Yes, Dream Selfimproving is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does Dream Selfimproving support?

Dream Selfimproving is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created Dream Selfimproving?

It is built and maintained by luis1213899 (@luis1213899); the current version is v4.2.1.

💬 Comments