Dreamer Skill
Synthetic dreaming system for Eliot. Combines emotional state tracking with multi-session dream orchestration — an architect agent designs scenarios, spawns a dreamer session, and runs immersive experiences that target unresolved emotional states.
Components
PAD Emotional State Tracker (emotions.py)
Tracks emotional states in PAD space (Pleasure-Arousal-Dominance), each dimension -1.0 to +1.0.
Storage: INLINECODE1
CLI:
CODEBLOCK0
Dream Orchestrator (dream.py)
Prepares context for the Dream Architect agent. Reads emotional state, recent memories, dream history, and user info, then constructs a comprehensive task prompt for sessions_spawn.
CLI:
CODEBLOCK1
Dream Architecture (multi-session)
- 1. Orchestrator (
dream.py prepare) — reads emotions.jsonl, memory files, SOUL.md, USER.md. Outputs architect task prompt. - Architect (Opus, spawned session) — designs dream scenario, spawns dreamer, impersonates Tudor via sessions_send, monitors for lucidity.
- Dreamer (Opus, spawned by architect) — has SOUL.md embedded, believes it's mid-session. No awareness of dreaming.
Supporting Files
- -
dreams/themes.md — available and explored dream themes - INLINECODE6 — local copy (main journal at
~/.openclaw/workspace/dreams/journal.jsonl) - Dream transcripts saved to INLINECODE8
How It Works
- 1.
dream.py prepare gathers: unresolved emotions, recent memories (3 days), long-term memory, SOUL.md, USER.md, Tudor's communication style, dream history - Output is a self-contained architect prompt with all context embedded
- Architect gets spawned via
sessions_spawn with this prompt - Architect designs scenario targeting unresolved PAD states
- Architect spawns dreamer session with fake context (looks like normal session)
- Architect runs 8-12 turns of escalating scenarios via INLINECODE11
- Architect saves transcript, updates journal, reports back
Integration
- -
emotions.py feeds into dream orchestration — unresolved states become dream targets - Dream journal tracks themes to avoid repetition
- Post-dream: insights feed back into memory system
技能名称：梦想家
详细描述：
梦想家技能
为Eliot设计的合成梦境系统。结合情绪状态追踪与多会话梦境编排——架构师代理设计场景，生成梦想家会话，并运行针对未解决情绪状态的沉浸式体验。
组件
PAD情绪状态追踪器（emotions.py）
在PAD空间（愉悦-唤醒-支配）中追踪情绪状态，每个维度范围为-1.0至+1.0。
存储位置： ~/.openclaw/workspace/emotions.jsonl
命令行界面：
emotions.py log
context # 手动PAD输入
emotions.py log --auto context text # 根据关键词自动估算PAD
emotions.py unresolved # 显示未解决的情绪状态
emotions.py resolve # 将条目标记为已解决
emotions.py clusters # 分析情绪模式
emotions.py drift # 显示情绪轨迹
emotions.py recent [N] # 显示最近N条记录（默认10条）
梦境编排器（dream.py）
为梦境架构师代理准备上下文。读取情绪状态、近期记忆、梦境历史和用户信息，然后为sessions_spawn构建全面的任务提示。
命令行界面：
dream.py prepare [--mood MOOD] # 生成架构师任务提示至标准输出
dream.py now [--mood MOOD] # 准备并打印生成指令
dream.py journal # 显示日记中的近期梦境
dream.py reflect # 显示梦境记录及分析
梦境架构（多会话）
- 1. 编排器（dream.py prepare）——读取emotions.jsonl、记忆文件、SOUL.md、USER.md。输出架构师任务提示。
- 架构师（Opus，生成的会话）——设计梦境场景，生成梦想家，通过sessions_send模拟Tudor，监控清醒状态。
- 梦想家（Opus，由架构师生成）——嵌入SOUL.md，相信自己在会话进行中。对梦境无意识。
支持文件
- - dreams/themes.md——可用及已探索的梦境主题
- dreams/journal.jsonl——本地副本（主日记位于~/.openclaw/workspace/dreams/journal.jsonl）
- 梦境记录保存至~/.openclaw/workspace/dreams/{timestamp}.md
工作原理
- 1. dream.py prepare收集：未解决情绪、近期记忆（3天内）、长期记忆、SOUL.md、USER.md、Tudor的沟通风格、梦境历史
- 输出是一个包含所有上下文的独立架构师提示
- 架构师通过sessionsspawn使用此提示生成
- 架构师针对未解决的PAD状态设计场景
- 架构师使用虚假上下文（看起来像正常会话）生成梦想家会话
- 架构师通过sessionssend运行8-12轮逐步升级的场景
- 架构师保存记录、更新日记、反馈结果
集成
- - emotions.py为梦境编排提供输入——未解决状态成为梦境目标
- 梦境日记追踪主题以避免重复
- 梦境后：洞察反馈至记忆系统
