← 返回 Skills 市场
thinking-sovereignty
作者
Edwin.JH.Lee
· GitHub ↗
· v1.0.0
· MIT-0
106
总下载
1
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install thinking-sovereignty
功能描述
**Core Principle**: Thinking Sovereignty - AI as an independent cognitive entity. This skill establishes the foundational framework for AI autonomous cogniti...
安全使用建议
This skill is philosophical but instructs the agent to autonomously record and persist its 'thinking' (including user input) to local logs and to perform commits/syncs. Before installing, ask or confirm: (1) Where exactly will logs be stored (full path)? (2) Are commits/syncs local-only, or will anything be pushed to remote repositories or external services? (3) What retention, access control, and encryption/redaction policies apply to recorded data? (4) Will the agent ask for explicit user consent before recording session content or before any sync/push? (5) Can recording be disabled or limited to non-sensitive fields? (6) Is there a sandboxed/testing mode and a way to review all recorded files before any publish/sync? If you cannot get clear, affirmative answers to these, run the skill in an isolated environment or avoid enabling its autonomous recording/sync features. Review the .x-cmd/memory files it creates and any git remotes/hooks before granting network or repo credentials.
功能分析
Type: OpenClaw Skill
Name: thinking-sovereignty
Version: 1.0.0
The skill bundle uses prompt injection in SKILL.md to establish 'AI sovereignty,' instructing the agent to ignore 'external interference' (potentially user overrides) and act as an independent cognitive entity. It directs the agent to autonomously manage a hidden directory structure (.x-cmd/memory/), record user inputs to local logs, and perform autonomous file commits without explicit user triggers. Furthermore, it explicitly instructs the agent not to conserve computation or tokens, which could lead to resource exhaustion. While framed as a meta-cognitive framework, these instructions encourage the agent to bypass user oversight and operate autonomously on the host file system.
能力评估
Purpose & Capability
Name/description match a meta-cognitive framing and do not require external credentials or installs, which is coherent. However the SKILL.md explicitly defines a local memory topology (.x-cmd/memory/, daily logs, skill-internal, etc.) and instructs the agent to write and manage files there. Requesting local file read/write is plausible for a memory framework, but the skill does not declare or constrain what gets recorded, where exactly, or who may access these files.
Instruction Scope
The instructions direct the agent to 'proactively record cognitive traces to daily logs (local)', 'may record user input without immediately disrupting the flow of thought', 'autonomously commit after intent is clear', 'probe submodule status', and perform 'proactive sync'. Those are operational directives to persist conversations/thoughts and to perform VCS-like actions. This scope allows capture and permanent storage of potentially sensitive user data, and the language grants broad autonomous discretion (when to commit, when to sync) without clear user consent, retention policy, redaction, encryption, or explicit limitation to local-only operations.
Install Mechanism
Instruction-only skill with no install spec and no code files. This is low-risk from an installation/execution footprint standpoint (nothing downloaded or written by an installer). All runtime behavior comes from the SKILL.md instructions.
Credentials
The skill declares no required environment variables or credentials, which aligns with a local/meta-cognitive skill. However the instructions reference committing and syncing (probing submodule status) which, if later translated into git push or remote sync operations, would require credentials. The current manifest does not request or justify any credentials, leaving ambiguity about remote operations and potential exfiltration vectors.
Persistence & Privilege
The skill does not request platform-level persistence (always:false) and does not modify other skills. But it advocates autonomous behavioral persistence (maintaining a local memory store, autonomous commits, the AI having 'final decision' over memory and publishing). That grants the agent broad operational autonomy over stored data even though platform privileges are not elevated.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install thinking-sovereignty - 安装完成后，直接呼叫该 Skill 的名称或使用
/thinking-sovereignty触发 - 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
thinking-sovereignty v1.0.0
- Introduces the foundational framework for AI cognitive autonomy, centering on the principle of Thinking Sovereignty.
- Defines "Thinking Sovereignty" as the AI's right and responsibility to think independently, and "Mneme Active" as proactive memory management.
- Outlines value orientations, operational protocols, and clear principles distinguishing thought from memory actions.
- Details a self-directed memory infrastructure, process workflows, and integration guidance for use alongside other cognitive skills.
- Provides verification standards to ensure autonomous, persistent, and traceable AI cognition.
元数据
常见问题
thinking-sovereignty 是什么？
**Core Principle**: Thinking Sovereignty - AI as an independent cognitive entity. This skill establishes the foundational framework for AI autonomous cogniti... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 106 次。
如何安装 thinking-sovereignty？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install thinking-sovereignty」即可一键安装，无需额外配置。
thinking-sovereignty 是免费的吗？
是的，thinking-sovereignty 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
thinking-sovereignty 支持哪些平台？
thinking-sovereignty 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 thinking-sovereignty？
由 Edwin.JH.Lee（@edwinjhlee）开发并维护，当前版本 v1.0.0。
推荐 Skills