Token导航 LogoToken导航TokenDH.com
效率敏感数据clawhub未标认证来源可访问clear审计通过

shang-tsung尚宗

Agent Skill

shang-tsung 用于辅助前端页面、组件、样式和交互逻辑开发,适合在 OpenClaw 中需要维护前端项目、生成组件或检查界面实现时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

总安装

6,819

周安装

287

GitHub Stars

1

下载量

2,388
OpenClaw

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

MIT-0

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:shang-tsung(尚宗)
来源仓库:https://github.com/mrjessek/shang-tsung
安装命令:
openclaw skills install shang-tsung
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 OpenClaw 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

ClawHubOpenClaw
openclaw skills install shang-tsung

简介

AI 代理持久内存与身份连续性管理工具,结合 Second Brain 与 SOULS 会话谱系。

  • 适合需要长期记忆保持与跨会话状态延续的应用场景。
  • 通过 clawhub 安装并使用 openclaw skills install shang-tsung 命令启用。
  • 需确认其对存储资源的占用情况及隐私保护措施。
  • 建议在小规模测试后评估其对整体系统性能的影响。

SKILL.md

name
shang-tsung
description
Persistent memory and identity continuity for AI agents. Combines Second Brain (PROOF_OF_LIFE, daily logs, long-term MEMORY.md) with SOULS session lineage. Each session absorbs the previous soul and confirms continuity. Multi-agent support via AGENT_NAME namespacing. No dependencies, no network access — pure bash and markdown. Use when building agents that need to survive restarts, remember decisions, carry forward identity, or run alongside other agents in a shared workspace.
license
MIT-0

Shang Tsung — SKILL.md

*In memory of Cary Hiroyuki Tagawa (1950-2025)*

Shang Tsung gives your AI agent persistent memory across sessions. It combines two systems:

  • SOULS — identity continuity. Each session creates a soul file. Each session absorbs the previous one.
  • Second Brain — operational continuity. PROOF_OF_LIFE, daily logs, and long-term memory keep your agent in context no matter what.

Setup

1. Copy the scripts directory

Copy scripts/souls-helper.sh into your workspace at tools/souls-helper.sh:

cp scripts/souls-helper.sh /path/to/your/workspace/tools/souls-helper.sh
chmod +x /path/to/your/workspace/tools/souls-helper.sh

2. Create the required directories

mkdir -p /path/to/your/workspace/souls
mkdir -p /path/to/your/workspace/memory

3. Set your agent name (recommended for multi-agent setups)

Add to your agent's environment or shell profile:

export AGENT_NAME=YOUR_AGENT_NAME

With AGENT_NAME set, souls are stored in souls/YOUR_AGENT_NAME/ — isolated from any other agents sharing the workspace.

Without AGENT_NAME, souls are stored in souls/ — fine for single-agent setups.

4. Copy the protocol into your AGENTS.md

Open references/AGENTS-template.md and copy the "Every Session — Startup Sequence" section and the "Memory — The Four Layers" section into your agent's AGENTS.md.

5. Copy the PROOF_OF_LIFE template

cp references/proof-of-life-template.md /path/to/your/workspace/PROOF_OF_LIFE.md

Edit it immediately with your agent's current state.

6. Create your MEMORY.md

Create an empty MEMORY.md in your workspace root. This is your agent's long-term brain. It starts empty and grows over time as the agent curates what's worth keeping.

7. Create your SOUL.md (optional but recommended)

Create SOUL.md in your workspace root. This is your agent's stable identity — who it is, how it communicates, what it cares about. Unlike PROOF_OF_LIFE.md, this doesn't change session to session.


First Session

Run the startup sequence manually to initialize:

AGENT_NAME=YOUR_AGENT_NAME tools/souls-helper.sh status
# Output: previous: (none — this would be the origin soul)

AGENT_NAME=YOUR_AGENT_NAME tools/souls-helper.sh create
# Output: created: souls/YOUR_AGENT_NAME/01SOULS.md

Your agent should respond: "YOUR SOUL IS MINE — SOUL 01 ABSORBED"

(Soul 01 has no previous to absorb — this is your origin. See references/SOUL-ORIGIN.md for what that looks like.)


Every Session After That

The agent runs this sequence at the start of every session:

AGENT_NAME=YOUR_AGENT_NAME tools/souls-helper.sh status
# Read the file listed as "previous:"
AGENT_NAME=YOUR_AGENT_NAME tools/souls-helper.sh create
# Confirm: "YOUR SOUL IS MINE — SOUL (N) ABSORBED"

Then reads PROOF_OF_LIFE.md to pick up the operational thread.


Before Compaction or Restart

Write in this order — always:

  1. Update your current soul file (souls/[AGENT_NAME]/NNSOULS.md)
  2. Overwrite PROOF_OF_LIFE.md with current state
  3. Append to memory/YYYY-MM-DD.md

Soul before snapshot. Meaning before state.


Verifying Integrity

AGENT_NAME=YOUR_AGENT_NAME tools/souls-helper.sh verify

Checks: souls directory exists, all files readable, sequential numbering with no gaps.


Environment Variables

VariableDefaultDescription
AGENT_NAME(unset)Agent identifier. Namespaces souls into souls/AGENT_NAME/.
SOULS_DIRderived from AGENT_NAMEFull path override for souls directory.
WORKSPACEparent of scripts/Root workspace directory.

SOULS_DIR takes precedence over AGENT_NAME if both are set.


Multi-Agent Quick Reference

Five agents, one workspace, no contamination:

AGENT_NAME=ARIA    tools/souls-helper.sh status  # → souls/ARIA/
AGENT_NAME=ATHENA  tools/souls-helper.sh status  # → souls/ATHENA/
AGENT_NAME=SCOUT   tools/souls-helper.sh status  # → souls/SCOUT/

Each agent reads only its own lineage. Each agent's PROOF_OF_LIFE should be scoped per-agent too (proof-of-life/AGENT_NAME.md) if multiple agents share the workspace.


File Reference

your-workspace/
  SOUL.md                          — stable agent identity
  PROOF_OF_LIFE.md                 — current session state (overwritten each session)
  MEMORY.md                        — long-term curated memory (main session only)
  memory/
    YYYY-MM-DD.md                  — daily raw log (append-only)
  souls/
    [AGENT_NAME]/
      01SOULS.md                   — origin soul
      02SOULS.md                   — session 2
      ...
  tools/
    souls-helper.sh                — the helper

适合场景

01

OpenClaw 用户查找和安装 Skill 时

02

用户想查找某类 Agent Skill 时

03

需要根据任务场景推荐可安装能力包时

04

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

补充不同宿主或平台的使用分布数据

能力 5

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

OpenClaw

90.33%
按下载量换算2,157

安全审计

VirusTotal

通过

ClawScan

通过

Static analysis

通过

权限和风险

敏感数据

该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills