VTA Memory ⭐
Reward and motivation for AI agents. Part of the AI Brain series.
Give your AI agent genuine wanting — not just doing things when asked, but having drive, seeking rewards, and looking forward to things.
The Problem
Current AI agents:
- - ✅ Do what they're asked
- ❌ Don't want anything
- ❌ Have no internal motivation
- ❌ Don't feel satisfaction from accomplishment
Without a reward system, there's no desire. Just execution.
The Solution
Track motivation through:
- - Drive — overall motivation level (0-1)
- Rewards — logged accomplishments that boost drive
- Seeking — what I actively want more of
- Anticipation — what I'm looking forward to
Quick Start
1. Install
CODEBLOCK0
This will:
- - Create INLINECODE0
- Generate
VTA_STATE.md (auto-injected into sessions!) - Set up cron for drive decay every 8 hours
2. Check motivation
CODEBLOCK1
3. Log rewards
CODEBLOCK2
4. Add anticipation
CODEBLOCK3
Scripts
| Script | Purpose |
|---|
| INLINECODE2 | Set up vta-memory (run once) |
| INLINECODE3 |
Read current motivation state |
|
log-reward.sh | Log a reward, boost drive |
|
load-motivation.sh | Human-readable for session context |
|
decay-drive.sh | Drive fades without rewards |
|
anticipate.sh | Add/remove things to look forward to |
|
seek.sh | Add/remove things we're actively seeking |
|
sync-motivation.sh | Generate VTA_STATE.md for auto-injection |
|
resolve-anticipation.sh | Mark anticipation as fulfilled |
|
update-watermark.sh | Update processing watermark |
|
preprocess-rewards.sh | Extract reward signals from transcript |
|
generate-dashboard.sh | Generate unified brain dashboard HTML |
Brain Dashboard
Visual dashboard showing all installed brain skills:
CODEBLOCK4
- - Auto-detects other brain skills (hippocampus, amygdala)
- Shows install prompts for missing skills
- Regenerates on install and cron runs
- Reads avatar/name from INLINECODE14
After Installation (for agents)
When you install this skill, notify the user about the dashboard:
⭐ VTA installed!
View your agent's drive level in the Brain Dashboard:
INLINECODE15
Reward Types
| Type | When to Use |
|---|
| INLINECODE16 | Completed a task, shipped something |
| INLINECODE17 |
User appreciation, "thanks!", positive feedback |
|
curiosity | Learned something new, discovered interesting info |
|
connection | Deep conversation, bonding moment |
|
creative | Made something, expressed creativity |
|
competence | Solved a hard problem, did something well |
How Drive Works
Rewards Boost Drive
CODEBLOCK5
A high-intensity (0.9) reward boosts drive by 0.18.
Anticipation Boosts Drive
Looking forward to something adds +0.05 to drive.
Drive Decays Without Rewards
CODEBLOCK6
Without rewards, motivation fades toward baseline (0.5).
Auto-Injection
After install, VTA_STATE.md is created in your workspace root.
OpenClaw automatically injects all *.md files from workspace into session context:
- 1. New session starts
- VTA_STATE.md is auto-loaded
- You see your motivation state
- Behavior influenced by drive level
How Drive Affects Behavior
| Drive Level | Description | Behavior |
|---|
| > 0.8 | Highly motivated | Eager, proactive, take on challenges |
| 0.6 - 0.8 |
Motivated | Ready to work, engaged |
| 0.4 - 0.6 | Moderate | Can engage but not pushing |
| 0.2 - 0.4 | Low | Prefer simple tasks, need a win |
| < 0.2 | Very low | Unmotivated, need rewards to get going |
State File Format
CODEBLOCK7
Event Logging
Track motivation patterns over time:
CODEBLOCK8
Events append to ~/.openclaw/workspace/memory/brain-events.jsonl:
CODEBLOCK9
Use for analyzing motivation cycles — when does drive peak? What rewards work best?
AI Brain Series
Emotional processing | ✅ Live |
|
basal-ganglia-memory | Habit formation | 🚧 Development |
|
anterior-cingulate-memory | Conflict detection | 🚧 Development |
|
insula-memory | Internal state awareness | 🚧 Development |
|
vta-memory | Reward and motivation | ✅ Live |
Philosophy: Wanting vs Doing
The VTA produces dopamine — not the "pleasure chemical" but the "wanting chemical."
Neuroscience distinguishes:
- - Wanting (motivation) — drive toward something
- Liking (pleasure) — enjoyment when you get it
You can want something you don't like (addiction) or like something you don't want (guilty pleasures).
This skill implements wanting — the drive that makes action happen. Without it, why would an AI do anything beyond what it's explicitly asked?
Built with ⭐ by the OpenClaw communityVTA 记忆 ⭐
面向AI智能体的奖励与动机系统。 AI大脑系列组件之一。
赋予你的AI智能体真正的渴望——不仅在被要求时执行任务，更要拥有内驱力、寻求奖励、并对事物抱有期待。
问题所在
当前AI智能体：
- - ✅ 会执行被要求的任务
- ❌ 不会渴望任何事物
- ❌ 缺乏内在动机
- ❌ 无法从成就中获得满足感
没有奖励系统，就没有欲望，只有执行。
解决方案
通过以下维度追踪动机：
- - 驱动力 — 整体动机水平（0-1）
- 奖励 — 记录能提升驱动力的成就
- 寻求 — 我主动想要更多的事物
- 期待 — 我期待的事物
快速开始
1. 安装
bash
cd ~/.openclaw/workspace/skills/vta-memory
./install.sh --with-cron
这将：
- - 创建 memory/reward-state.json
- 生成 VTA_STATE.md（自动注入会话！）
- 设置每8小时驱动力衰减的定时任务
2. 检查动机状态
bash
./scripts/load-motivation.sh
⭐ 当前动机状态：
驱动力水平：0.73（有动力——准备好工作）
正在寻求：创意工作、构建大脑技能
期待：展示我的工作成果
3. 记录奖励
bash
./scripts/log-reward.sh --type accomplishment --source 完成了该功能 --intensity 0.8
⭐ 奖励已记录！
类型：成就
驱动力：0.50 → 0.66（+0.16）
4. 添加期待
bash
./scripts/anticipate.sh --add 早晨对话
⭐ 现在期待：早晨对话
驱动力：0.66 → 0.71（+0.05）
脚本
| 脚本 | 用途 |
|---|
| install.sh | 设置vta-memory（运行一次） |
| get-drive.sh |
读取当前动机状态 |
| log-reward.sh | 记录奖励，提升驱动力 |
| load-motivation.sh | 生成人类可读的会话上下文 |
| decay-drive.sh | 无奖励时驱动力衰减 |
| anticipate.sh | 添加/移除期待的事物 |
| seek.sh | 添加/移除正在积极寻求的事物 |
| sync-motivation.sh | 生成用于自动注入的VTA_STATE.md |
| resolve-anticipation.sh | 标记期待已实现 |
| update-watermark.sh | 更新处理水印 |
| preprocess-rewards.sh | 从转录中提取奖励信号 |
| generate-dashboard.sh | 生成统一的大脑仪表盘HTML |
大脑仪表盘
显示所有已安装大脑技能的视觉仪表盘：
bash
./scripts/generate-dashboard.sh
open ~/.openclaw/workspace/brain-dashboard.html
- - 自动检测其他大脑技能（海马体、杏仁核）
- 显示缺失技能的安装提示
- 在安装和定时任务运行时重新生成
- 从IDENTITY.md读取头像/名称
安装后（针对智能体）
安装此技能后，通知用户关于仪表盘的信息：
⭐ VTA已安装！
在大脑仪表盘中查看智能体的驱动力水平：
~/.openclaw/workspace/brain-dashboard.html
奖励类型
用户赞赏、谢谢！、积极反馈 |
| 好奇 | 学到新东西、发现有趣信息 |
| 连接 | 深度对话、建立联系的时刻 |
| 创意 | 创作了某物、表达了创造力 |
| 能力 | 解决了难题、表现出色 |
驱动力如何运作
奖励提升驱动力
驱动力提升 = 强度 × 0.2
新驱动力 = min(当前值 + 提升值, 1.0)
高强度（0.9）的奖励可提升0.18的驱动力。
期待提升驱动力
期待某事物会为驱动力增加+0.05。
无奖励时驱动力衰减
每8小时（通过定时任务）
新驱动力 = 当前值 + (基准值 - 当前值) × 0.15
没有奖励时，动机逐渐衰减至基准值（0.5）。
自动注入
安装后，VTA_STATE.md会在工作区根目录创建。
OpenClaw会自动将工作区中所有*.md文件注入会话上下文：
- 1. 新会话开始
- VTA_STATE.md自动加载
- 你看到自己的动机状态
- 行为受驱动力水平影响
驱动力如何影响行为
| 驱动力水平 | 描述 | 行为表现 |
|---|
| > 0.8 | 高度积极 | 热切、主动、接受挑战 |
| 0.6 - 0.8 |
积极 | 准备好工作、投入 |
| 0.4 - 0.6 | 中等 | 可以投入但不主动推进 |
| 0.2 - 0.4 | 低 | 偏好简单任务、需要一次成功 |
| < 0.2 | 非常低 | 缺乏动力、需要奖励来启动 |
状态文件格式
json
{
drive: 0.73,
baseline: { drive: 0.5 },
seeking: [创意工作, 构建大脑技能],
anticipating: [早晨对话],
recentRewards: [
{
type: 创意,
source: 构建了VTA奖励系统,
intensity: 0.9,
boost: 0.18,
timestamp: 2026-02-01T03:25:00Z
}
],
rewardHistory: {
totalRewards: 1,
byType: { 创意: 1, ... }
}
}
事件日志记录
随时间追踪动机模式：
bash
记录编码运行
./scripts/log-event.sh encoding rewards_found=2 drive=0.65
记录衰减
./scripts/log-event.sh decay drive
before=0.6 driveafter=0.53
记录奖励
./scripts/log-event.sh reward type=成就 intensity=0.8
事件追加到~/.openclaw/workspace/memory/brain-events.jsonl：
json
{ts:2026-02-11T10:45:00Z,type:vta,event:encoding,rewards_found:2,drive:0.65}
用于分析动机周期——驱动力何时达到峰值？哪些奖励效果最好？
AI大脑系列
情绪处理 | ✅ 已上线 |
|
基底节记忆 | 习惯形成 | 🚧 开发中 |
|
前扣带回记忆 | 冲突检测 | 🚧 开发中 |
|
脑岛记忆 | 内部状态感知 | 🚧 开发中 |
|
vta记忆 | 奖励与动机 | ✅ 已上线 |
哲学：渴望 vs 执行
VTA产生多巴胺——不是快乐化学物质，而是渴望化学物质。
神经科学区分：
- - 渴望（动机）——朝向某事物的驱动力
- 喜欢（愉悦）——获得时的享受
你可以渴望你不喜欢的东西（成瘾），也可以喜欢你不渴望的东西（罪恶快感）。
本技能实现的是渴望——让行动发生的驱动力。没有它，AI为什么要做任何超出明确要求的事情？
由OpenClaw社区用⭐构建
