Xpersona
agentCLAWHUBUnverified

Agent Memory System v12

Agent 记忆系统 — 五路融合检索 + 双时间线 + 因果链 + Spirit管家 + 记忆回声 + 弹性配置 + Circuit Breaker + GDPR合规 + 192项安全审计修复 Skill: Agent Memory System v12 Owner: concisegjh Summary: Agent 记忆系统 — 五路融合检索 + 双时间线 + 因果链 + Spirit管家 + 记忆回声 + 弹性配置 + Circuit Breaker + GDPR合规 + 192项安全审计修复 Tags: agent:4.0.0, ai:4.0.0, latest:12.2.0, memory:4.0.0, rag:4.0.0, sqlite:4.0.0, vector:4.0.0 Version history: v12.2.0 | 2026-06-08T12:52:25.840Z | user **Agent Memory 12.2.0 – Major update with enhanced documentation, security, and OpenClaw integration** - Refa

OpenClaw
View on ClawHubHomepage

Rank

62

Safety

84

Downloads

2.4k

Updated

Oct 9, 2026

Version

12.2.0

Source

CLAWHUB

About

What it does, and when to use it.

Capability contract not published. No trust telemetry is available yet. 2.4K downloads reported by the source. Last updated 10/9/2026.

Avoid when

  • Contract metadata is missing or unavailable for deterministic execution.

Risk flags: missing_or_unavailable_contract, trust_data_unavailable, schema_references_missing

Public facts

Every fact links back to the source it came from.

Vendor
Clawhubvendor · observed Oct 9, 2026
Protocol compatibility
OpenClawcompatibility · observed Oct 9, 2026
Adoption signal
2.4K downloadsadoption · observed Oct 9, 2026
Latest release
12.2.0release · observed Jun 8, 2026
Handshake status
UNKNOWNsecurity

Install and run

Setup complexity: low.

clawhub skill install s170csdzwc049dc4ympczmb2v9844bz7:super-memory
  1. Setup complexity is classified as HIGH. You must provision dedicated cloud infrastructure or an isolated VM. Do not run this directly on your local workstation.
  2. Final validation: Expose the agent to a mock request payload inside a sandbox and trace the network egress before allowing access to real customer data.

Contract: missing

curl -s "https://www.xpersona.co/api/v1/agents/clawhub-concisegjh-super-memory/snapshot"

Documentation

CLAWHUB

160,000 characters of source documentation, loaded on request.

Extracted files

5 files captured from the source.

SKILL.md

---
name: agent-memory
version: 12.2.0
description: Agent 记忆系统 — 五路融合检索 + 双时间线 + 因果链 + Spirit管家 + 记忆回声 + 弹性配置 + Circuit Breaker + GDPR合规 + 192项安全审计修复
author: Agent Memory Contributors
license: MIT
source: https://github.com/agent-memory/agent-memory
tags:
  - memory
  - agent
  - rag
  - vector
  - sqlite
  - semantic-search
  - multi-agent
  - multimodal
  - distillation
  - fts5
  - circuit-breaker
  - gdpr
  - security
homepage: https://github.com/agent-memory/agent-memory
repository: https://github.com/agent-memory/agent-memory
user-invocable: true
install:
  type: pip
  spec: requirements.txt
  python: ">=3.10"
  steps:
    - "pip install -r requirements.txt"
  provenance:
    core: vendored (pure Python + stdlib, no external build)
    optional:
      - name: sqlite-vec
        source: https://pypi.org/project/sqlite-vec/
        pinned: ">=0.1.9"
      - name: sentence-transformers
        source: https://pypi.org/project/sentence-transformers/
        pinned: ">=2.2.0"
      - name: FlagEmbedding
        source: https://pypi.org/project/FlagEmbedding/
        pinned: ">=1.2.0"
    models:
      - name: BAAI/bge-small-zh-v1.5
        source: https://huggingface.co/BAAI/bge-small-zh-v1.5
        verify: SHA256 after first download recommended
credentials:
  required: []
  optional:
    - name: AGENT_MEMORY_API_KEY
      description: HTTP 服务 API Key（生产部署必须，CLI 模式不需要）
      env: AGENT_MEMORY_API_KEY
    - name: AGENT_MEMORY_ADMIN_PASSWORD
      description: 多用户模式管理员密码（仅 PermissionManager 使用，CLI 模式不需要）
      env: AGENT_MEMORY_ADMIN_PASSWORD
    - name: AGENT_MEMORY_API_KEY_READ
      description: 只读角色 API Key（HTTP 服务可选）
      env: AGENT_MEMORY_API_KEY_READ
    - name: AGENT_MEMORY_API_KEY_WRITE
      description: 读写角色 API Key（HTTP 服务可选）
      env: AGENT_MEMORY_API_KEY_WRITE
    - name: OPENAI_API_KEY
      description: OpenAI embedding/LLM 后端（可选，不设置则使用本地模式）
      env: OPENAI_API_KEY
    - name: SILICONFLOW_API_KEY
      description: SiliconFlow LLM 后端（可选）
      env: SILICONFLOW_API_KEY
    - name: COHERE_API_KEY
      description: Cohere embedding 后端（可选）
      env: COHERE_API_KEY
    - name: VOYAGE_API_KEY
      description: Voyage embedding 后端（可选）
      env: VOYAGE_API_KEY
    - name: CUSTOM_LLM_API_KEY
      description: 自定义 LLM 后端（可选）
      env: CUSTOM_LLM_API_KEY
    - name: CUSTOM_LLM_BASE_URL
      description: 自定义 LLM 地址（可选）
      env: CUSTOM_LLM_BASE_URL
permissions:
  required:
    - name: filesystem
      description: "读写本地SQLite数据库和备份文件（~/.agent_memory/）"
      scope: "read/write: ~/.agent_memory/**"
    - name: network
      description: "HTTP API服务和远程LLM/Embedding调用"
      scope: "outbound: https; inbound: configurable port (default 8988)"
    - name: subprocess
      description: "模型守护进程管理（model_server.py start/stop）和媒体处理（ffmpeg）"
      scope: "python3, ffmpeg, ffprobe only"
  optional:
    - name: clipboard
      description: "分享卡片复制（需用户确

agent_memory/README.md

# Agent Memory V12 — Internal Developer Guide

> **Version**: 12.0.0 | **Python**: ≥3.10 | **License**: MIT

This is the **internal** developer guide for the `agent_memory` package.
For user-facing documentation, see the [root README](../README.md) and [ARCHITECTURE.md](../ARCHITECTURE.md).

## What Is This Package?

Agent Memory is a **personal memory operating system** for AI agents.
It goes beyond passive storage — the **Spirit butler** proactively patrols, diagnoses, and repairs the memory store, while four specialized engines handle ingestion, retrieval, maintenance, and cognition.

## V12 Architecture Summary

### TEMPR Five-Lane Retrieval

Five parallel retrieval lanes fused via Reciprocal Rank Fusion (RRF, k=60):

| Lane | Method | Module |
|------|--------|--------|
| 1 | FTS5 + LIKE (structured full-text) | `fts_manager.py`, `store.py` |
| 1.5 | BM25 sparse retrieval | `bm25_index.py` |
| 2 | Semantic vector (vec0 / ChromaDB) | `embedding_store.py` |
| 3 | Entity expansion (relationship graph) | `entity.py` |
| 4 | Causal chain traversal (max depth 2) | `memory_links` via `recall_engine.py` |

### Four Engines

| Engine | File | Responsibility |
|--------|------|----------------|
| IngestEngine | `engines/ingest.py` | Dedup, quality scoring, auto-linking |
| RecallEngine | `engines/recall_engine.py` | TEMPR 5-lane retrieval + RRF + intent + MMR |
| MaintainEngine | `engines/maintain.py` | Lifecycle, decay, distillation |
| CognitionEngine | `engines/cognition.py` | Causal reasoning, knowledge graph |

### 13 Mixins

`AgentMemory` composes behavior via mixins in `mixins/`:

MemoryMixin · RecallMixin · SessionMixin · MaintenanceMixin · DistillMixin · EncyclopediaMixin · TimelineMixin · StatsMixin · PersonaMixin · RoleMixin · MediaStyleMixin · ReactorMixin · ExportMixin

### Spirit Butler

`spirit/` — Dual-LLM security protocol, natural language command interface, proactive health checks, cross-agent dedup, preference sync, memory partitioning. Isolated from core via `spirit/interface.py`.

## Module Directory

| Subpackage / Module | Description |
|----------------------|-------------|
| `engines/` | Four core engines + decay policy + curiosity + federation + feedback + metacognitive loop |
| `mixins/` | 13 behavior mixins composed into AgentMemory |
| `spirit/` | Spirit butler: commands, health checks, reports, LLM layer, proactive delivery |
| `storage/` | SQLite store, embedding store, CryptoStore, FTS manager, agent manager, cache |
| `config/` | Unified settings (env > JSON > defaults), schema.sql, dimension definitions |
| `collectors/` | Multi-source collectors: DingTalk, WeChat, Email, Calendar, File + scheduler |
| `enterprise/` | Audit log, compliance guard, knowledge distiller, offboarding, permission matrix, skill marketplace |
| `privacy/` | PII analyzer, consent management, guard, patterns, rules |
| `plugins/` | Plugin framework: auto-tagger, sentiment monitor, Obsidian sync, Slack notifier |
| `integration/` | LangChain co

README.md

# 🧠 Agent Memory v12.2.0

> Give your AI agents a memory that thinks — not just stores.

[![Security Audit](https://img.shields.io/badge/security-192%20fixes-brightgreen)](https://github.com/agent-memory/agent-memory/security) [![GDPR](https://img.shields.io/badge/compliance-GDPR-blue)](https://github.com/agent-memory/agent-memory) [![Circuit Breaker](https://img.shields.io/badge/resilience-circuit%20breaker-orange)](https://github.com/agent-memory/agent-memory)

Agent Memory is a production-grade memory layer for AI agents. It goes beyond simple vector storage with **5-way retrieval**, **temporal reasoning**, **proactive management**, and **enterprise-grade privacy**.

## ✨ Why Agent Memory?

| Feature | Agent Memory | Mem0 | Zep | LangChain Memory |
|---------|-------------|------|-----|------------------|
| 5-way retrieval (FTS+BM25+Semantic+Entity+Causal) | ✅ | ❌ | ❌ | ❌ |
| Dual-timeline fact management | ✅ | ❌ | ❌ | ❌ |
| Proactive Spirit butler | ✅ | ❌ | ❌ | ❌ |
| Chinese native support (12 PII types) | ✅ | ❌ | ❌ | ❌ |
| Enterprise compliance (GDPR + SOC2) | ✅ | ❌ | ❌ | ❌ |
| Self-hosted / Docker | ✅ | ✅ | ✅ | ✅ |

## 🚀 Quick Start

### Install

```bash
pip install agent-memory
```

### 5-Minute Demo

```python
from agent_memory import Memory

# Initialize — zero config
mem = Memory()

# Remember
result = mem.remember("Alice is a data scientist working on NLP projects")
print(result.memory_id)  # mem_abc123

# Recall
results = mem.recall("NLP projects")
for r in results.items:
    print(r["content"])

# Forget
mem.forget(result.memory_id)  # Soft delete, restorable within 30 days

# PII Detection & Redaction
from agent_memory.privacy.guard import PrivacyGuard
guard = PrivacyGuard()
redacted = guard.redact("Contact: 13812345678")
# → "Contact: [手机号]"

# Health Check
info = mem.status()
print(f"Healthy: {info.get('healthy')} | Total: {info.get('total_memories')}")
```

### Docker

```bash
docker compose up -d
# API available at http://localhost:8988
```

### Playground

```bash
pip install "agent-memory[web]"
python -m agent_memory.playground.app
# Open http://localhost:8988
```

## 🏗 Architecture

```
┌─────────────────────────────────────────────┐
│  Adapters: SDK / REST API / MCP / CLI       │
├─────────────────────────────────────────────┤
│  Spirit: Proactive Butler + Safety Protocol │
├─────────────────────────────────────────────┤
│  Cognition: Digital Twin + Metacognition    │
├─────────────────────────────────────────────┤
│  Engines: Ingest / Recall / Maintain / Graph│
├─────────────────────────────────────────────┤
│  Storage: SQLite + FTS5 + sqlite-vec       │
└─────────────────────────────────────────────┘
```

## 📦 Installation Options

```bash
# Core (minimal dependencies)
pip install agent-memory

# With web API
pip install "agent-memory[web]"

# With semantic search
pip install "agent-memory[semantic]"

# With Chinese NLP
pip install "agent-memory[chinese]"

# With observability
pip install "agent-memory[observability]"

_meta.json

{
  "ownerId": "kn74z18dmw2dhgfyxxmnfr28h584403j",
  "slug": "super-memory",
  "version": "12.2.0",
  "publishedAt": 1780923145840
}

agent_memory/API.md

# Agent Memory API Reference

完整 API 参考：HTTP 端点 + Python 接口 + 各模块详细说明。

> 快速接入和架构概览请见 [README.md](README.md)。

---

## API 版本策略

本项目提供三个 HTTP API 版本，各版本状态和定位如下：

| 版本 | 入口文件 | 端口 | 状态 | 说明 |
|------|---------|------|------|------|
| **v1** | `server.py` | 8976 | **stable** | 纯 stdlib HTTP 服务，零外部依赖，适合轻量部署和开发调试 |
| **v2** | `api_v2.py` | 8978 | **deprecated** | FastAPI + Pydantic，RFC 7807 错误格式，不再积极维护，建议迁移至 v3 |
| **v3** | `api_v3.py` | 8988 | **active** | FastAPI 异步 + SSE + 多租户隔离 + JWT/API Key 双认证，推荐生产使用 |

### 版本选择建议

- **新项目**：直接使用 **v3**（`api_v3.py`），获得多租户、异步、SSE 等完整功能
- **轻量/嵌入式**：使用 **v1**（`server.py`），无 FastAPI 依赖，单文件部署
- **v2 用户**：请迁移至 v3，v2 仅保留向后兼容，不再接受新功能

### 启动命令

```bash
# v1 (stable)
python3 server.py --api-key mysecret

# v2 (deprecated)
uvicorn agent_memory.api_v2:app --host 127.0.0.1 --port 8978

# v3 (active, 推荐)
uvicorn agent_memory.api_v3:app --host 0.0.0.0 --port 8988
```

---

## 错误码汇总

### HTTP 状态码

| HTTP 状态码 | 含义 | 触发场景 |
|------------|------|---------|
| 200 | 成功 | 所有成功请求 |
| 201 | 创建成功 | `POST /remember` 写入记忆 |
| 204 | 无内容 | 删除操作成功 |
| 400 | 请求错误 | 参数缺失/格式错误/内容超长 |
| 401 | 未认证 | 缺少认证头或 Token 无效（v2/v3） |
| 403 | 权限不足 | API Key 角色不匹配、Agent ID 覆盖被拒、强制认证端点未提供 Key |
| 404 | 未找到 | 路径不存在、记忆 ID 不存在 |
| 429 | 请求过多 | 超出速率限制（v2/v3 token bucket） |
| 500 | 服务器内部错误 | 未预期异常 |
| 503 | 服务不可用 | SSE 客户端数达上限、服务过载 |

### 业务错误码（v2/v3 RFC 7807 格式）

| 错误码 | HTTP 状态 | 说明 |
|--------|----------|------|
| `AUTH_MISSING` | 401 | 缺少认证信息，需提供 `Authorization: Bearer <token>` 或 `X-API-Key` |
| `AUTH_INVALID` | 401 | 认证凭证无效（Token 签名错误 / API Key 不存在） |
| `TOKEN_EXPIRED` | 401 | JWT Token 已过期 |
| `INSUFFICIENT_PERMISSIONS` | 403 | 当前角色无权执行此操作 |
| `FORBIDDEN` | 403 | 访问被拒绝（Agent ID 绑定冲突等） |

### v1 (server.py) 错误消息

| 错误消息 | HTTP 状态 | 说明 |
|---------|----------|------|
| `unauthorized` | 401 | 缺少或无效的 API Key |
| `content is required` | 400 | 写入请求缺少 content 字段 |
| `content too long: N chars (max 50000)` | 400 | 单条内容超长 |
| `query is required` | 400 | recall 请求缺少 query |
| `query too long: N chars (max 10000)` | 400 | 查询超长 |
| `messages array is required` | 400 | batch 请求缺少 messages |
| `too many messages: N (max 50)` | 400 | 批量写入超限 |
| `memory_id is required` | 400 | feedback/update 缺少 memory_id |
| `Write endpoint requires API Key authentication` | 403 | 写入端点强制要求 API Key |
| `Insufficient permissions (role: X)` | 403 | 角色权限不足 |
| `Agent ID override not authorized` | 403 | API Key 绑定的 Agent ID 与请求不一致 |
| `SSE client limit reached` | 503 | SSE 连接数达上限 |
| `memory not found or no version history` | 404 | 记忆不存在 |

### Python 异常层级

| 异常类 | 父类 | 说明 |
|--------|------|------|
| `MemoryError` | `Exception` | 所有 agent_memory 异常的基类 |
| `StorageError` | `MemoryError` | 数据库/存储操作失败 |
| `MemoryNotFoundError` | `MemoryError` | 请求的 memory_id 不存在 |
| `DuplicateMemoryError` | `MemoryError` | 重复内容/哈希冲突 |
| `FilterRejectedError` | `MemoryError` | 内容被过滤器拒绝 |
| `DeduplicationError` | `Memory

Related agents

Same protocol, similar job.

Github ReposUpdated 2h agoRank 70

AionUi

Free, local, open-source 24/7 Cowork app and OpenClaw for Gemini CLI, Claude Code, Codex, OpenCode, Qwen Code, Goose CLI, Auggie, and more | 🌟 Star if you like it!

MCPOPENCLAW
Github ReposUpdated 6mo agoRank 70

activepieces

AI Agents & MCPs & AI Workflow Automation • (~400 MCP servers for AI agents) • AI Automation / AI Agent with MCPs • AI Workflows & AI Agents • MCPs for AI Agents

OPENCLAW
Github ReposUpdated 6mo agoRank 70

cherry-studio

AI productivity studio with smart chat, autonomous agents, and 300+ assistants.

MCPOPENCLAW
Github ReposUpdated 7mo agoRank 70

CopilotKit

The Frontend for Agents & Generative UI. React + Angular

OPENCLAW

Machine-readable data

The same record, as JSON, for agents and crawlers.

{
  "facts": [
    {
      "factKey": "vendor",
      "category": "vendor",
      "label": "Vendor",
      "value": "Clawhub",
      "href": "https://clawhub.ai/concisegjh/skills/super-memory",
      "sourceUrl": "https://clawhub.ai/concisegjh/skills/super-memory",
      "sourceType": "profile",
      "confidence": "medium",
      "observedAt": "2026-10-09T15:14:28.516Z",
      "isPublic": true
    },
    {
      "factKey": "protocols",
      "category": "compatibility",
      "label": "Protocol compatibility",
      "value": "OpenClaw",
      "href": "https://www.xpersona.co/api/v1/agents/clawhub-concisegjh-super-memory/contract",
      "sourceUrl": "https://www.xpersona.co/api/v1/agents/clawhub-concisegjh-super-memory/contract",
      "sourceType": "contract",
      "confidence": "medium",
      "observedAt": "2026-10-09T15:14:28.516Z",
      "isPublic": true
    },
    {
      "factKey": "traction",
      "category": "adoption",
      "label": "Adoption signal",
      "value": "2.4K downloads",
      "href": "https://clawhub.ai/concisegjh/super-memory",
      "sourceUrl": "https://clawhub.ai/concisegjh/super-memory",
      "sourceType": "profile",
      "confidence": "medium",
      "observedAt": "2026-10-09T15:14:28.516Z",
      "isPublic": true
    },
    {
      "factKey": "latest_release",
      "category": "release",
      "label": "Latest release",
      "value": "12.2.0",
      "href": "https://clawhub.ai/concisegjh/super-memory",
      "sourceUrl": "https://clawhub.ai/concisegjh/super-memory",
      "sourceType": "release",
      "confidence": "medium",
      "observedAt": "2026-06-08T12:52:25.840Z",
      "isPublic": true
    },
    {
      "factKey": "handshake_status",
      "category": "security",
      "label": "Handshake status",
      "value": "UNKNOWN",
      "href": "https://www.xpersona.co/api/v1/agents/clawhub-concisegjh-super-memory/trust",
      "sourceUrl": "https://www.xpersona.co/api/v1/agents/clawhub-concisegjh-super-memory/trust",
      "sourceType": "trust",
      "confidence": "medium",
      "observedAt": null,
      "isPublic": true
    }
  ],
  "events": [
    {
      "eventType": "release",
      "title": "Release 12.2.0",
      "description": "**Agent Memory 12.2.0 – Major update with enhanced documentation, security, and OpenClaw integration** - Refactored documentation system: added ARCHITECTURE.md, CONTRIBUTING.md, SECURITY.md, UPGRADE_PLAN.md, and extensive internal docs. - CLI migration: all commands now use `cli.py` with adjusted instructions for OpenClaw skill directory usage. - Credentials expanded: now supports granular read/write API keys, new LLM/embedding backend options, and improved configuration guidance. - Added explicit permissions section with risk assessment for sensitive operations and optional external integrations. - Many legacy and example files removed or archived to streamline the codebase. - Numerous security, compliance, and operational improvements reflected in updated instructions and environment variable handling.",
      "href": "https://clawhub.ai/concisegjh/super-memory",
      "sourceUrl": "https://clawhub.ai/concisegjh/super-memory",
      "sourceType": "release",
      "confidence": "medium",
      "observedAt": "2026-06-08T12:52:25.840Z",
      "isPublic": true
    }
  ]
}

Record generated Oct 9, 2026.

Sponsored

Ads related to Agent Memory System v12 and adjacent AI workflows.