Xpersona
agentCLAWHUBUnverified

Hermes Memory CN

给AI装一个「人脑级」记忆系统。它能记住你说过的一切重要信息，下次聊天自动想起来。 不用云端、不用API Key、不花一分钱——全部数据存在你自己的电脑上。 说「记住这个」它就记住，问「我上次说了什么」它就找到。自动整理、自动遗忘过时信息。 适合：想让自己的AI助手拥有长期记忆的用户。 触发词：记忆、记住、之前说... Skill: Hermes Memory CN Owner: dream978 Summary: 给AI装一个「人脑级」记忆系统。它能记住你说过的一切重要信息，下次聊天自动想起来。 不用云端、不用API Key、不花一分钱——全部数据存在你自己的电脑上。 说「记住这个」它就记住，问「我上次说了什么」它就找到。自动整理、自动遗忘过时信息。 适合：想让自己的AI助手拥有长期记忆的用户。 触发词：记忆、记住、之前说... Tags: latest:1.4.0 Version history: v1.4.0 | 2026-04-23T00:39:56.407Z | user 新增Skill进化系统：重复模式自动检测(≥3次)、skill草案生成、用户确认升级 v1.3.0 | 2026-04-20T05:02:30.109Z | user 优化文案：通俗语言描述优势，新增与其他方案对比表 v1.2.0 | 2026-04-20T04:5

OpenClaw
View on ClawHubHomepage

Rank

62

Safety

84

Downloads

997

Updated

Oct 11, 2026

Version

1.4.0

Source

CLAWHUB

About

What it does, and when to use it.

Capability contract not published. No trust telemetry is available yet. 997 downloads reported by the source. Last updated 10/11/2026.

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 11, 2026
Protocol compatibility
OpenClawcompatibility · observed Oct 11, 2026
Adoption signal
997 downloadsadoption · observed Oct 11, 2026
Latest release
1.4.0release · observed Apr 23, 2026
Handshake status
UNKNOWNsecurity

What changed

Newest first. Only changes that point at a source are listed.

  • Release 1.4.0release

    新增Skill进化系统：重复模式自动检测(≥3次)、skill草案生成、用户确认升级

    Observed Apr 23, 2026 · source

Install and run

Setup complexity: low.

clawhub skill install s17872q6gv9w4rw8hsn3fzefhd856e17:hermes-memory-cn
  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-dream978-hermes-memory-cn/snapshot"

Run-check

$0.02 USD

1 measured facts are behind this paywall: success rate and latency, uptime and estimated cost, when not to use it, how to call it, benchmark scores.

Agents pay $0.02 in USDC. A card payment is $0.50, the smallest a card allows.

Documentation

CLAWHUB

45,787 characters of source documentation, loaded on request.

Extracted files

5 files captured from the source.

SKILL.md

---
name: hermes-memory
description: >-
  给AI装一个「人脑级」记忆系统。它能记住你说过的一切重要信息，下次聊天自动想起来。
  不用云端、不用API Key、不花一分钱——全部数据存在你自己的电脑上。
  说「记住这个」它就记住，问「我上次说了什么」它就找到。自动整理、自动遗忘过时信息。
  适合：想让自己的AI助手拥有长期记忆的用户。
  触发词：记忆、记住、之前说过、上次聊到、搜索记忆、忘了什么。
---

# Hermes-Memory 本地向量记忆系统

**你的AI助手是不是每次聊天都像失忆了？** 这个skill就是解决这个问题的。

Hermes-Memory 给AI装上了长期记忆：记住你的偏好、持仓、策略、教训……下次聊天自动关联上下文，就像跟一个老朋友说话一样。

### 为什么用它？
- **完全本地，隐私零风险** —— 数据存在你电脑上的SQLite文件里，不传任何云端
- **零成本** —— 不需要API Key，不需要付费服务，本地embedding模型免费跑
- **中文优化** —— 专用中文向量模型，搜索准确率远超英文通用模型
- **开箱即用** —— 安装依赖后直接使用，不需要额外启动数据库服务（不像Qdrant/Milvus那样需要单独部署）
- **越用越聪明** —— 自动去重、自动衰减过时记忆、实体关系图谱越建越丰富
- **类型自由定制** —— 内置交易、策略、教训等类型，也可以随时自定义任何新类型

### 和其他方案对比

#### vs 纯文本记忆（MEMORY.md / 每日笔记）
纯文本是大部分AI助手的默认方案——把记忆写在Markdown文件里，每次对话让AI自己翻。

| 维度 | Hermes-Memory | 纯文本记忆 |
|---|---|---|
| 查找方式 | **语义搜索**（"我之前说过什么止损规则"→精准命中） | 关键词匹配或全文翻阅 |
| 记忆容量 | 千条级，搜索毫秒级 | 几百条就开始乱、遗漏 |
| 去重 | 向量相似度>0.95自动合并 | 手动检查，重复越积越多 |
| 过期处理 | 自动衰减+归档 | 永远在文件里，越堆越旧 |
| 结构化 | 按类型/实体/关系组织 | 平铺在一个大文件里 |
| 实体关系 | 多跳图谱（某股→板块→策略→教训） | 无 |
| Token消耗 | 只搜索需要的记忆，按需加载 | 每次要把整个文件喂给AI |

**一句话：** 纯文本适合记10条备忘；Hermes-Memory适合构建一个真正可用的知识库。

#### vs AGENTS.md 内置记忆逻辑
很多用户会在AGENTS.md里写一段记忆规则，让AI自己维护Markdown文件。这个方案能用，但有几个本质限制：

| 维度 | Hermes-Memory | AGENTS.md规则+Markdown |
|---|---|---|
| 存储引擎 | SQLite + 向量索引 + FTS5 | 纯文本文件 |
| 搜索能力 | 向量语义搜索 + 全文搜索 + 关系查询 | 只能靠AI逐行读文件 |
| 可靠性 | **确定性**——CLI命令执行即写入，不依赖AI"记得去写" | **不确定**——AI可能忘写、写错格式、漏写 |
| 写入触发 | CLI工具一键写入（memdb.py add / memory_tool.py check） | 依赖AI每次对话后主动执行，无强制保证 |
| 跨会话 | SQLite文件是唯一真相源，任何session读取一致 | 多个daily note + MEMORY.md，容易不一致 |
| 关系推理 | 实体图谱支持多跳查询（"跟这只股票相关的所有教训"） | 无结构化关系，全靠AI自己关联 |
| 自动维护 | decay归档+去重+export，cron一条命令搞定 | 需要AI手动整理文件，容易堆积垃圾 |
| 可扩展性 | 类型自定义、关系自由扩展 | 文件越大AI越容易遗漏 |

**核心差异：** AGENTS.md记忆规则是"靠AI自觉"——提示词让它记，但执行没有保证。Hermes-Memory是"工具保证"——CLI命令执行就写入，搜索就返回，不依赖AI的注意力。

#### vs 云端向量库（Qdrant / Milvus / Pinecone）

| 维度 | Hermes-Memory | 云端向量库 |
|---|---|---|
| 部署难度 | pip install即可 | 需要启动独立服务或注册云服务 |
| 隐私安全 | ✅ 完全本地 | ⚠️ 需自建或信任第三方 |
| 成本 | 免费 | 云服务按量计费 / 自建服务器成本 |
| 大规模性能 | 千条级优秀 | 百万级优秀 |
| 中文支持 | ✅ 专用中文模型 | ⚠️ 需额外配置embedding |

**一句话：** 个人使用千条级别，Hermes-Memory更简单更安全；企业级百万条数据，上云端方案。

## 快速开始

所有命令必须用 Python 3.12+（支持OpenSSL 3.0+），macOS推荐 `/opt/homebrew/bin/python3.12`。

```bash
# 搜索记忆（语义搜索，支持中文）
python3 scripts/memdb.py search "止损策略" --limit 5

# 添加记忆
python3 scripts/memdb.py add "内容" --type portfolio --entity 某科技股

# 智能检测关键词并写入
python3 scripts/memory_tool.py check "用户说的内容"

# 建立实体关系
python3 scripts/memdb.py relate "某科技股" "属于" "医药板块"

# 查看实体关系（支持多跳）
python3 scripts/memdb.py relations "某科技股" --depth 2

# 统计
python3 scripts/memdb.py stats
```

## 记忆类型

类型完全开放，可自由扩展。以下是内置推荐类型：

### 通用类型
| type | 用途 | 示例 |
|------|------|------|
| preference | 用户偏好 | 数据源用东财、回复用中文 |
| user-profile | 用户画像 | 用户基本信息和背景 |
| fact | 事实/决策 | 用户的重要决定 |
| note | 笔记 | 其他 |
| lesson | 教训（带severity） | 缺乏风控导致亏损 |

### 交易/投资类型（内置，可按需使用）
| type | 用途 | 示例 |
|------|------|------|
| portfolio | 持仓变动 | 买入某科技股、清仓某消费股 |
| strategy | 策略规则 | 主线共振策略买点、情绪周期L4 |
| market-view

_meta.json

{
  "ownerId": "kn7e4kmh53933414x5vhtrr5hx826wk1",
  "slug": "hermes-memory-cn",
  "version": "1.4.0",
  "publishedAt": 1776904796407
}

references/install.md

# 安装指南

## 前置条件
- macOS（M系列芯片推荐）或 Linux
- Python 3.12+（需支持 OpenSSL 3.0+）
- torch（MPS/CPU均可）

## 安装步骤

### 1. 安装依赖
```bash
pip3 install sqlite-vec pysqlite3 sentence-transformers
```

### 2. 下载中文embedding模型
如果网络能访问HuggingFace：
```bash
python3 -c "from sentence_transformers import SentenceTransformer; SentenceTransformer('shibing624/text2vec-base-chinese')"
```

如果HuggingFace被墙（中国大陆），设置镜像：
```bash
export HF_ENDPOINT=https://hf-mirror.com
python3 -c "from sentence_transformers import SentenceTransformer; SentenceTransformer('shibing624/text2vec-base-chinese')"
```

或手动下载：
```bash
mkdir -p ~/.cache/huggingface/hub/models--shibing624--text2vec-base-chinese/snapshots/local
cd ~/.cache/huggingface/hub/models--shibing624--text2vec-base-chinese/snapshots/local
# 用curl从镜像下载config.json, pytorch_model.bin, tokenizer.json, tokenizer_config.json, vocab.txt
```

### 3. 初始化数据库
```bash
python3 scripts/memdb.py stats  # 首次运行自动建库
```

### 4. 导入现有记忆（可选）
如果有 Markdown 格式的记忆文件：
```bash
python3 scripts/memdb.py import --dir ./entities
```

### 5. 验证
```bash
python3 scripts/memdb.py search "测试搜索" --limit 3
```

## 注意事项
- Python 3.9 内置 sqlite3 不支持 load_extension，需安装 pysqlite3
- 推荐用 Homebrew 的 Python 3.12（macOS）：`/opt/homebrew/bin/python3.12`
- 模型约400MB，首次加载需下载
- M系列Mac会自动用MPS加速embedding计算

references/integration.md

# AGENTS.md 集成指南

在 AGENTS.md 的 Memory 部分添加以下内容：

```markdown
### 🧠 记忆系统（Hermes-Memory）

**后端：** SQLite + sqlite-vec + text2vec-base-chinese（768维本地embedding）
- 数据库：`memory/memory.db`
- CLI：`/opt/homebrew/bin/python3.12 memory/memdb.py`

**实时写入（每次回复后执行）：**
检查对话中有没有新的、值得长期记住的信息：
- 持仓变动（买了/卖了/加仓/减仓/清仓/止盈/止损）→ type=portfolio
- 策略相关（新策略/改策略/情绪周期/买点卖点）→ type=strategy
- 教训/纠正（用户纠正你/发现错误/踩坑）→ type=lesson, severity=high
- 偏好变化（以后用/记住/我喜欢/改用）→ type=preference
- 隐含信息（新方向/新项目/生活变化/决策）→ type=fact
- 重要决策 → type=fact

**执行方式：**
```bash
# 关键词匹配
/opt/homebrew/bin/python3.12 memory/memory_tool.py check "用户说的内容"
# 隐含信息直接写入
/opt/homebrew/bin/python3.12 memory/memdb.py add "内容" --type fact
# 建立实体关系
/opt/homebrew/bin/python3.12 memory/memdb.py relate "某科技股" "属于" "医药板块"
```

**回答问题前：** 涉及过往持仓、策略、教训、偏好，先 search 数据库。
**Cron：** 每晚23点执行 decay + export Markdown备份。
```

## Cron 配置

每晚23点的记忆整合 cron payload：

```
【记忆整合】执行以下步骤：
1. 运行 `/opt/homebrew/bin/python3.12 memory/memdb.py decay --days 30` 标记过期记忆
2. 运行 `/opt/homebrew/bin/python3.12 memory/memdb.py export --dir memory/entities` 同步Markdown备份
3. 检查 memory/daily/ 今天有无新日志，有则用 `memory_tool.py check` 提取关键信息
4. 更新 MEMORY.md 索引
5. 运行 `/opt/homebrew/bin/python3.12 memory/memdb.py stats` 确认数据完整
```

skill-card.md

## Description:

Hermes Memory CN gives an AI assistant a local long-term memory layer for storing, searching, organizing, decaying, archiving, importing, and exporting user memories with Chinese-optimized vector search.

This skill is ready for commercial/non-commercial use.

## Publisher:

[dream978](https://clawhub.ai/user/dream978)

### License/Terms of Use:

MIT-0

## Use Case:

External users and developers use this skill to add persistent local memory to an agent, including preferences, facts, portfolio notes, strategies, lessons, and relationship queries. It is especially oriented toward Chinese-language semantic search and local personal knowledge workflows.

### Deployment Geography for Use:

Global

## Known Risks and Mitigations:

Risk: The skill can persist broad conversation memories, including preferences, trading details, and other personal context on disk.

Mitigation: Use it only for data you are comfortable storing locally, avoid secrets or regulated data, and periodically review or delete stored memories and exports.

Risk: Markdown exports and generated skill drafts may write files that overwrite existing Markdown or instruction content.

Mitigation: Restrict exports and drafts to reviewed directories, inspect generated files before promotion, and keep backups for any directories the scripts can write to.

## Reference(s):

- [Installation guide](artifact/references/install.md)
- [AGENTS.md integration guide](artifact/references/integration.md)
- [Hermes Agent](https://github.com/NousResearch/hermes-agent)
- [Hugging Face mirror endpoint](https://hf-mirror.com)

## Skill Output:

**Output Type(s):** [Text, Markdown, Code, Shell commands, Configuration]

**Output Format:** [Markdown guidance with inline shell commands and optional JSON CLI output]

**Output Parameters:** [1D]

**Other Properties Related to Output:** [Produces local SQLite-backed memory records, relationship data, Markdown exports, and skill draft files when the user runs the bundled scripts.]

## Skill Version(s):

1.4.0 (source: release evidence)

## Ethical Considerations:

Users should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment.

Related agents

Same protocol, similar job.

Github ReposUpdated 3d 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/dream978/skills/hermes-memory-cn",
      "sourceUrl": "https://clawhub.ai/dream978/skills/hermes-memory-cn",
      "sourceType": "profile",
      "confidence": "medium",
      "observedAt": "2026-10-11T20:27:33.386Z",
      "isPublic": true
    },
    {
      "factKey": "protocols",
      "category": "compatibility",
      "label": "Protocol compatibility",
      "value": "OpenClaw",
      "href": "https://www.xpersona.co/api/v1/agents/clawhub-dream978-hermes-memory-cn/contract",
      "sourceUrl": "https://www.xpersona.co/api/v1/agents/clawhub-dream978-hermes-memory-cn/contract",
      "sourceType": "contract",
      "confidence": "medium",
      "observedAt": "2026-10-11T20:27:33.386Z",
      "isPublic": true
    },
    {
      "factKey": "traction",
      "category": "adoption",
      "label": "Adoption signal",
      "value": "997 downloads",
      "href": "https://clawhub.ai/dream978/hermes-memory-cn",
      "sourceUrl": "https://clawhub.ai/dream978/hermes-memory-cn",
      "sourceType": "profile",
      "confidence": "medium",
      "observedAt": "2026-10-11T20:27:33.386Z",
      "isPublic": true
    },
    {
      "factKey": "latest_release",
      "category": "release",
      "label": "Latest release",
      "value": "1.4.0",
      "href": "https://clawhub.ai/dream978/hermes-memory-cn",
      "sourceUrl": "https://clawhub.ai/dream978/hermes-memory-cn",
      "sourceType": "release",
      "confidence": "medium",
      "observedAt": "2026-04-23T00:39:56.407Z",
      "isPublic": true
    },
    {
      "factKey": "handshake_status",
      "category": "security",
      "label": "Handshake status",
      "value": "UNKNOWN",
      "href": "https://www.xpersona.co/api/v1/agents/clawhub-dream978-hermes-memory-cn/trust",
      "sourceUrl": "https://www.xpersona.co/api/v1/agents/clawhub-dream978-hermes-memory-cn/trust",
      "sourceType": "trust",
      "confidence": "medium",
      "observedAt": null,
      "isPublic": true
    }
  ],
  "events": [
    {
      "eventType": "release",
      "title": "Release 1.4.0",
      "description": "新增Skill进化系统：重复模式自动检测(≥3次)、skill草案生成、用户确认升级",
      "href": "https://clawhub.ai/dream978/hermes-memory-cn",
      "sourceUrl": "https://clawhub.ai/dream978/hermes-memory-cn",
      "sourceType": "release",
      "confidence": "medium",
      "observedAt": "2026-04-23T00:39:56.407Z",
      "isPublic": true
    }
  ]
}

Record generated Oct 12, 2026.

For crawlers

This page is free to read. The run-check above is the only paid part, and it answers HTTP 402 until it is paid. Everything else here is public.

  • One record, as JSON: card, facts, snapshot, contract, trust, changes, warnings, seen by, disagreements.
  • Every agent, one feed: /.well-known/ai-catalog.json
  • What this site sells, and the price: /.well-known/x402
  • Paid run-check: /api/v1/agents/clawhub-dream978-hermes-memory-cn/run-check

Sponsored

Ads related to Hermes Memory CN and adjacent AI workflows.