← 返回 Skills 市场
ar3ss12

Soul architect

作者 Ar3ss12 · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ⚠ suspicious
218
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install soul-architect-v1
功能描述
Call when creating/updating STYLE_MANIFESTO persona source code.
安全使用建议
This skill appears to do what it says (generate STYLE_MANIFESTO files), but its mandatory research steps encourage collecting highly sensitive or potentially defamatory personal information. Before installing or using: (1) avoid using 'personal' mode on living private individuals without consent; prefer 'avatar' or 'legend' where plausible; (2) review and sanitize SKILL.md to remove any nonprintable/control characters and any implicit instruction that could leak data or override safety rules; (3) restrict the agent's browsing/data sources (or disable web access) if you don't want scraping of private material; (4) inspect any knowledge_base.json files placed under personas/ to ensure they contain only allowed data; (5) monitor outputs and logs for disclosure of PII or unverified claims; and (6) if uncertain, run the skill only in an isolated sandbox and review generated manifestos before deploying them in production.
功能分析
Type: OpenClaw Skill Name: soul-architect-v1 Version: 1.0.0 The skill bundle is designed for creating detailed AI persona 'manifestos' for roleplay or content generation. The SKILL.md file provides a structured research protocol and a template for character development, while scripts/synthesize.py acts as a simple coordinator that creates local directories and signals the agent to generate content. There is no evidence of data exfiltration, malicious execution, or harmful prompt injection; the 'God Tier' terminology is stylistic rather than functional, and the .gitignore file correctly excludes potentially sensitive persona data.
能力评估
Purpose & Capability
The name, description, SKILL.md manifesto template, and the lightweight scripts/synthesize.py align: this is a persona/style-manifesto authoring helper. No unrelated binaries, env vars, or external services are requested.
Instruction Scope
SKILL.md mandates running at least six targeted web-research queries that seek sensitive material (childhood trauma, 'psychological diagnosis rumors', 'treatment of subordinates', losing-control compilations, etc.). That scope can prompt collection of private, stigmatizing, or defamatory information about real people. The template is prescriptive and forces a single structure; it also includes a lightweight protocol string (MANIFESTO_PENDING) to trigger agent synthesis. Additionally, a pre-scan flagged 'unicode-control-chars' inside SKILL.md, which indicates possible hidden control characters or prompt-injection attempts embedded in the instructions.
Install Mechanism
No install spec and only a small Python script are included. There is no download/extract or external installer. Risk from installation is low.
Credentials
The skill requests no credentials or environment variables (proportionate). However, the 'personal' mode expects a local knowledge_base.json (private persona data) and the research protocol encourages fetching potentially sensitive external data; absence of declared credentials does not eliminate privacy/exfiltration risks via web searches or agent's integrated browsing abilities.
Persistence & Privilege
The skill is not always-enabled and does not request elevated platform privileges or attempt to modify other skills. It only creates persona directories/files under personas/{name}, which is within its stated scope.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install soul-architect-v1
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /soul-architect-v1 触发
  4. 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v1.0.0
soul-architect 1.0.0 initial release - Introduces "Soul Architect" skill for creating STYLE_MANIFESTO persona source code. - Provides a detailed Manifesto Template (v4.1) to define character identity, cognition, behavior, and linguistic style. - Mandates structured research protocol requiring at least 6 targeted queries across psychology, body language, worldview, and more. - Outlines strict usage, goal, validation, and output requirements to ensure quality and consistency. - Specifies prohibited shortcuts and mandates concrete deliverables for completion.
元数据
Slug soul-architect-v1
版本 1.0.0
许可证 MIT-0
累计安装 0
当前安装数 0
历史版本数 1
常见问题

Soul architect 是什么?

Call when creating/updating STYLE_MANIFESTO persona source code. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 218 次。

如何安装 Soul architect?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install soul-architect-v1」即可一键安装,无需额外配置。

Soul architect 是免费的吗?

是的,Soul architect 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

Soul architect 支持哪些平台?

Soul architect 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。

谁开发了 Soul architect?

由 Ar3ss12(@ar3ss12)开发并维护,当前版本 v1.0.0。

💬 留言讨论