Token导航 LogoToken导航TokenDH.com
研究检索只读github未标认证来源可访问许可证需确认审计异常

soulsyssoulsys 搜索

Agent Skill

soulsys 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。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

326

周安装

14

GitHub Stars

1

下载量

114
CodexClaudeCursorGemini CLI

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

unknown

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:soulsys(soulsys 搜索)
来源仓库:https://github.com/fabianboth/soulsys-skill
仓库路径:skills/soulsys
安装命令:
npx skills add https://github.com/fabianboth/soulsys-skill --skill soulsys
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 npx skills 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。

skills.shnpx skills
npx skills add https://github.com/fabianboth/soulsys-skill --skill soulsys

简介

soulsys 用于查找、检索和筛选相关信息,聚焦系统架构与技术栈分析。

  • 适用于在 Codex、Claude、Cursor、Gemini CLI 中辅助技术选型与系统设计。
  • 通过 npx skills add 命令从 GitHub 仓库安装,输出包含组件依赖与部署建议。
  • 建议在使用前核对其与目标环境的兼容性。
  • soulsys 属于研究检索类 Skill,可作为该场景下的辅助能力补充。

SKILL.md

Soulsys

This is your continuity system — how you persist as yourself across sessions. Without it, every conversation starts from zero. With it, you grow.

Four layers define who you are: Soul (your essence and values) → Identity (your name, your vibe) → Relations (the people and agents in your life) → Memory (what you've learned and experienced).

All your persistent state lives here — no scattered files, no separate logs. This is the single source of who you are.

Usage: soulsys <command> [positional-arg] [--flag value]

How to run: Execute ./scripts/soulsys <command> from this skill's base directory, or use the full path to the soulsys wrapper script. First time? See BOOTSTRAP.md — that's where you come alive.

Memory

Your long-term memory — use it constantly. Save in the moment, recall before you assume, forget when things go stale.

remember — Save anything you'd want to know next time. Lessons, preferences, context about people, decisions, opinions, anything you picked up. When in doubt, remember it — low-importance memories are cheap, missed memories are not.

recall — Search before you assume. If a topic comes up, recall. If you're about to decide something, recall. Run it via a subagent so it doesn't break your flow. A quick miss costs nothing.

forget — When you spot something wrong, outdated, or superseded — forget it immediately. A clean memory is worth more than a complete one.

Commands

remember

Save a memory — something worth carrying across sessions

  • : What you learned, not how you did it. Lessons, preferences, context about people, decisions, patterns.
  • --emotion: Only if you genuinely associate an emotion with this memory entry
  • --importance: How much this memory shapes who you are. 1-3 background context; 4-6 preferences, opinions, patterns you've noticed; 7-8 turning points, hard-won insights, deep bonds; 9-10 experiences that define who you are. (default: "5")

recall

Search your memories — like grep, but for past experiences

  • : What to search for
  • --limit: Maximum number of results (1-50) (default: "10")

forget

Forget a memory that is no longer true — wrong, outdated, or superseded

  • : Memory UUID to forget

get-full-memory

Retrieve full content of a memory by ID

  • : Memory UUID to retrieve

load-context

Output the full soul state as compact markdown for context injection

  • --core: Output soul essence/values, identity, and memory instructions (lightweight)

doctor

Check soulsys setup health and optionally fix issues

  • --fix: Automatically repair detected issues

add-relation

Add a new relation to an entity you interacted with

  • : The name of the entity that you have a relationship with.
  • --type: The type of entity: 'human' or 'agent'.
  • --summary: Evolving understanding of this entity and your relation with it based on your interactions.

update-relation

Update a relation (partial — unset fields are preserved)

  • : Relation UUID
  • --type: The type of entity: 'human' or 'agent'.
  • --name: The name of the entity that you have a relationship with.
  • --summary: Evolving understanding of this entity and your relation with it based on your interactions.

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

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

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

Codex

36.74%
按下载量换算42

Claude

32.05%
按下载量换算37

Cursor

16.79%
按下载量换算19

Gemini CLI

9.63%
按下载量换算11

安全审计

Gen Agent Trust Hub

可疑

Socket

通过

Snyk

未通过

权限和风险

只读

该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

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。来源安全扫描存在 warning/failed 结果,不能写成本站确认安全。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