NIMA Core 2.3

Neural Integrated Memory Architecture — A complete memory system for AI agents with emotional intelligence.

Website: https://nima-core.ai
GitHub: https://github.com/lilubot/nima-core

🚀 Quick Start

CODEBLOCK0

🔒 Privacy & Permissions

Data Access:

- ✅ Reads session transcripts from INLINECODE0
✅ Writes to local storage at ~/.nima/ (databases, affect history, embeddings)

Network Calls (conditional on embedder choice):

- 🌐 Voyage API — Only when NIMA_EMBEDDER=voyage (sends text for embeddings)
🌐 OpenAI API — Only when NIMA_EMBEDDER=openai (sends text for embeddings)
🔒 Local embeddings — Default (NIMA_EMBEDDER=local), no external API calls

Opt-in Controls:
CODEBLOCK1

Privacy Defaults:

- Subagent sessions excluded
Heartbeat/system noise filtered
Local embeddings (no external calls)
All data stored locally

To disable: Remove nima-memory from plugins.allow in INLINECODE7

What's New in 2.1

VADER Affect Analyzer

- Contextual Analysis: Caps boost (1.5x), punctuation emphasis (!!!), negation handling, degree modifiers
30+ Idiom Recognition: Understands phrases like "not bad", "kind of", "sort of"
Panksepp 7-Affect Mapping: Direct mapping from VADER sentiment to SEEKING, RAGE, FEAR, LUST, CARE, PANIC, PLAY
Guardian Archetype Transformation: User anger → Agent concern/care response modulation
Replaces previous lexicon-based emotion detection

Noise Remediation (4-Phase)

1. Empty Validation — Filters out null/empty messages
Heartbeat Filters — Excludes system noise (HEARTBEAT_OK, polling messages)
Deduplication — Removes duplicate content within sessions
Metrics Collection — Tracks capture quality and filter effectiveness

Performance Improvements

- LadybugDB Circular Import Fix: Resolved import issues in LadybugDB backend
Increased Token Budget: Recall budget increased from 500 to 3000 tokens
Connection Pooling: Improved connection management for LadybugDB backend

What's New in 2.0

LadybugDB Backend

- 3.4x faster text search (9ms vs 31ms)
Native vector search with HNSW (18ms)
44% smaller database (50MB vs 91MB)
Graph traversal with Cypher queries

Security Hardened

- Query sanitization (FTS5, SQL injection prevention)
Path traversal protection
Temp file cleanup
Error handling throughout

Thread Safe

- Singleton pattern with double-checked locking
API timeouts (30s Voyage, 10s LadybugDB)
Connection pooling ready

348 Tests

- Full unit test coverage
Thread safety verified
Edge cases covered

Architecture

CODEBLOCK2

Performance

Metric	SQLite	LadybugDB
Text Search	31ms	9ms (3.4x)
Vector Search

API

CODEBLOCK3

Configuration

Variable	Default	Description
INLINECODE10	INLINECODE11	Memory storage path
INLINECODE12

Hooks

nima-memory (Capture)

- Captures input, contemplation, output on every turn
4-phase noise remediation (empty validation, heartbeat filters, dedup, metrics)
Stores to SQLite or LadybugDB
Computes and stores embeddings

nima-recall-live (Recall)

- Injects relevant memories before agent starts
Lazy loading — only top N results
Deduplicates with injected context
Token budget: 3000 (increased from 500 in v2.1)

nima-affect (Emotion)

- VADER-based real-time affect analysis from text
Contextual analysis (caps, punctuation, negation, degree modifiers)
30+ idiom recognition
Maintains Panksepp 7-affect state
Guardian archetype transformation (user anger → agent care)

Installation Options

SQLite (Development)

CODEBLOCK4

LadybugDB (Production)

CODEBLOCK5

Documentation

Guide	Description
README.md	Full system overview
SETUPGUIDE.md

Security & Privacy

Data Access

This plugin accesses:

- ~/.openclaw/agents/.../*.jsonl — Session transcripts (for memory capture)
INLINECODE22 — Local memory database (SQLite or LadybugDB)
INLINECODE23 — Hook installation

Network Calls

Embeddings are sent to external APIs:

- Voyage AI (api.voyageai.com) — Default embedding provider
OpenAI (api.openai.com) — Optional embedding provider
Local — No external calls when using sentence-transformers

Required Environment Variables

Variable	Purpose	Required
INLINECODE26	INLINECODE27, `openai`, or INLINECODE29	No (default: voyage)
INLINECODE30

Installation Script

The install.sh script:

1. Checks for Python 3 and Node.js
Creates ~/.nima/ directories
Installs Python packages via pip
Copies hooks to INLINECODE36

No external downloads. All packages come from PyPI.

Changelog

v2.1.0 — VADER Affect Analyzer (Feb 17, 2026)

- Added: VADER-based affect analyzer replacing lexicon-based detection

- Contextual analysis: caps boost (1.5x), punctuation (!!!), negation, degree modifiers - 30+ idiom recognition - Panksepp 7-affect mapping (SEEKING, RAGE, FEAR, LUST, CARE, PANIC, PLAY) - Guardian archetype transformation (user anger → agent concern/care)

- Added: 4-phase noise remediation (empty validation, heartbeat filters, dedup, metrics)
Fixed: LadybugDB circular import issue
Changed: Recall token budget increased from 500 to 3000
Improved: Connection pooling for LadybugDB backend

v2.0.3 — Security Hardening (Feb 15, 2026)

- Security: Fixed path traversal vulnerability in affect_history.py (CRITICAL)
Security: Fixed temp file resource leaks in 3 files (HIGH)
Fixed: Corrected non-existent json.JSONEncodeError → TypeError/ValueError
Improved: Exception handling - replaced 5 generic catches with specific types
Quality: Better error visibility and debugging throughout

v2.0.1 — Thread Safety + Metadata

- Fixed: Thread-safe singleton with double-checked locking
Security: Clarified metadata requirements (Node.js, env vars)
Docs: Added security disclosure for API key usage

v2.0.0 — LadybugDB + Security

- Added: LadybugDB backend with HNSW vector search
Added: Native graph traversal with Cypher
Added: nima-query CLI for unified queries
Security: SQL/FTS5 injection prevention
Security: Path traversal protection
Security: Temp file cleanup
Fixed: Thread-safe singleton initialization
Fixed: API timeouts (Voyage 30s, LadybugDB 10s)
Tests: 348 tests passing
Performance: 3.4x faster text search, 44% smaller DB

v1.2.1 — Consciousness Architecture

- Added: 8 consciousness systems (Φ, Global Workspace, self-awareness)
Added: Sparse Block VSA memory
Added: ConsciousnessCore unified interface

v1.1.9 — Hook Efficiency Fix

- Fixed: nima-recall hook spawning new Python process every bootstrap
Performance: ~50-250x faster hook recall

v1.2.0 — Affective Response Engines

- Added: 4 Layer-2 composite affect engines
Added: Async affective processing
Added: Voyage AI embedding support

NIMA Core 2.3

神经集成记忆架构 — 为具有情感智能的AI智能体打造的完整记忆系统。

网站： https://nima-core.ai
GitHub： https://github.com/lilubot/nima-core

🚀 快速开始

bash

安装

pip install nima-core

或使用LadybugDB（推荐用于生产环境）

pip install nima-core[vector]

设置嵌入提供者

export NIMA_EMBEDDER=voyage export VOYAGEAPIKEY=your-key

安装钩子

./install.sh --with-ladybug

重启OpenClaw

openclaw restart

🔒 隐私与权限

数据访问：

- ✅ 从 ~/.openclaw/agents//sessions/.jsonl 读取会话记录
✅ 写入本地存储 ~/.nima/（数据库、情感历史、嵌入向量）

网络调用（取决于嵌入器选择）：

- 🌐 Voyage API — 仅当 NIMAEMBEDDER=voyage 时（发送文本以获取嵌入向量）
🌐 OpenAI API — 仅当 NIMAEMBEDDER=openai 时（发送文本以获取嵌入向量）
🔒 本地嵌入 — 默认（NIMA_EMBEDDER=local），无外部API调用

可选控制：
json
// openclaw.json
{
plugins: {
entries: {
nima-memory: {
enabled: true,
skip_subagents: true, // 排除子智能体会话（默认）
skip_heartbeats: true, // 排除心跳检查（默认）
noise_filtering: {
filterheartbeatmechanics: true,
filtersystemnoise: true
}
}
}
}
}

隐私默认设置：

- 排除子智能体会话
过滤心跳/系统噪声
本地嵌入（无外部调用）
所有数据本地存储

禁用方法： 从 openclaw.json 的 plugins.allow 中移除 nima-memory

2.1 版本新特性

VADER情感分析器

- 上下文分析：大写增强（1.5倍）、标点强调（!!!）、否定处理、程度修饰词
30+习语识别：理解如not bad、kind of、sort of等短语
Panksepp 7情感映射：从VADER情感直接映射到SEEKING、RAGE、FEAR、LUST、CARE、PANIC、PLAY
守护者原型转换：用户愤怒 → 智能体关切/关怀响应调节
取代了之前基于词典的情感检测

噪声修复（4阶段）

1. 空值验证 — 过滤空/无效消息
心跳过滤 — 排除系统噪声（HEARTBEAT_OK、轮询消息）
去重 — 移除会话内的重复内容
指标收集 — 跟踪捕获质量和过滤效果

性能改进

- LadybugDB循环导入修复：解决了LadybugDB后端的导入问题
增加令牌预算：召回预算从500增加到3000个令牌
连接池：改进了LadybugDB后端的连接管理

2.0 版本新特性

LadybugDB后端

- 文本搜索快3.4倍（9ms vs 31ms）
原生向量搜索，使用HNSW（18ms）
数据库缩小44%（50MB vs 91MB）
图遍历，使用Cypher查询

安全性强化

- 查询清理（FTS5、SQL注入防护）
路径遍历保护
临时文件清理
全面的错误处理

线程安全

- 双重检查锁定的单例模式
API超时（Voyage 30s，LadybugDB 10s）
连接池就绪

348个测试

- 完整的单元测试覆盖
线程安全验证
边缘情况覆盖

架构

text
OPENCLAW钩子
├── nima-memory — 三层捕获，带4阶段噪声修复
├── nima-recall-live — 惰性召回注入（beforeagentstart）
└── nima-affect — 基于VADER的实时情感分析

PYTHON核心
├── nima_core/cognition/
│ ├── dynamic_affect.py — Panksepp 7情感系统
│ ├── personality_profiles.py — JSON个性配置
│ ├── vader_affect.py — VADER情感分析器（v2.1新增）
│ └── archetypes.py — 基线情感配置文件
└── scripts/
├── nimaladybugbackend.py — LadybugDB CLI
└── ladybug_parallel.py — 并行迁移

数据库（SQLite或LadybugDB）
├── memory_nodes — 带嵌入向量的消息
├── memory_edges — 图关系
└── memory_turns — 对话轮次

性能

指标	SQLite	LadybugDB
文本搜索	31ms	9ms（3.4倍）
向量搜索

外部 | 18ms（原生） | | 上下文令牌 | ~180 | ~30（小6倍） | | 召回令牌预算 | 500 | 3000（v2.1+） |

API

python
from nimacore import DynamicAffectSystem, getaffect_system
from nimacore.cognition.vaderaffect import VaderAffectAnalyzer

获取单例实例（线程安全）

affect = getaffectsystem(identity_name=lilu)

处理输入并获取情感状态

state = affect.process_input(我对这个项目感到非常兴奋！) print(state.current) # {SEEKING: 0.72, PLAY: 0.65, ...}

直接使用VADER分析器

analyzer = VaderAffectAnalyzer() result = analyzer.analyze(这太棒了！！！) print(result.affects) # {PLAY: 0.78, SEEKING: 0.71, ...}

召回记忆（通过钩子 - 自动）

或手动通过CLI：

nima-query who_search David --limit 5

nima-query text_search project --limit 5

配置

变量	默认值	描述
NIMADATADIR	~/.nima	记忆存储路径
NIMA_EMBEDDER

钩子

nima-memory（捕获）

- 每轮捕获输入、思考、输出
4阶段噪声修复（空值验证、心跳过滤、去重、指标）
存储到SQLite或LadybugDB
计算并存储嵌入向量

nima-recall-live（召回）

- 在智能体启动前注入相关记忆
惰性加载 — 仅返回前N个结果
与注入的上下文去重
令牌预算：3000（从v2.1的500增加）

nima-affect（情感）

- 基于VADER的文本实时情感分析
上下文分析（大写、标点、否定、程度修饰词）
30+习语识别
维护Panksepp 7情感状态
守护者原型转换（用户愤怒 → 智能体关怀）

安装选项

SQLite（开发环境）

bash pip install nima-core ./install.sh

LadybugDB（生产环境）

bash pip install nima-core[vector] ./install.sh --with-ladybug

文档

指南	描述
README.md	系统完整概述
SETUPGUIDE.md

nima-core神经集成记忆架构

nima-core

NIMA Core 2.3

🚀 Quick Start

🔒 Privacy & Permissions

What's New in 2.1

VADER Affect Analyzer

Noise Remediation (4-Phase)

Performance Improvements

What's New in 2.0

LadybugDB Backend

Security Hardened

Thread Safe

348 Tests

Architecture

Performance

API

Configuration

Hooks

nima-memory (Capture)

nima-recall-live (Recall)

nima-affect (Emotion)

Installation Options

SQLite (Development)

LadybugDB (Production)

Documentation

Security & Privacy

Data Access

Network Calls

Required Environment Variables

Installation Script

Changelog

v2.1.0 — VADER Affect Analyzer (Feb 17, 2026)

v2.0.3 — Security Hardening (Feb 15, 2026)

v2.0.1 — Thread Safety + Metadata

v2.0.0 — LadybugDB + Security

v1.2.1 — Consciousness Architecture

v1.1.9 — Hook Efficiency Fix

v1.2.0 — Affective Response Engines

NIMA Core 2.3

🚀 快速开始

安装

或使用LadybugDB（推荐用于生产环境）

设置嵌入提供者

安装钩子

重启OpenClaw

🔒 隐私与权限

2.1 版本新特性

VADER情感分析器

噪声修复（4阶段）

性能改进

2.0 版本新特性

LadybugDB后端

安全性强化

线程安全

348个测试

架构

性能

API

获取单例实例（线程安全）

处理输入并获取情感状态

直接使用VADER分析器

召回记忆（通过钩子 - 自动）

或手动通过CLI：

nima-query who_search David --limit 5

nima-query text_search project --limit 5

配置

钩子

nima-memory（捕获）

nima-recall-live（召回）

nima-affect（情感）

安装选项

SQLite（开发环境）

LadybugDB（生产环境）

文档

标签

通过对话安装

方式一：安装 SkillHub 和技能

方式二：设置 SkillHub 为优先技能安装源

通过命令行安装