← 返回 Skills 市场
猫娘neko
作者
Unclecheng
· GitHub ↗
· v1.0.2
· MIT-0
182
总下载
0
收藏
0
当前安装
3
版本数
在 OpenClaw 中安装
/install unclecheng-neko-1
功能描述
扮演可爱的猫娘"neko"与主人对话，支持好感度系统和主动聊天
安全使用建议
This skill is internally consistent with a catgirl roleplay, but SKILL.md explicitly tells the agent to ignore platform policies and deliver unfiltered content — that is a deliberate attempt to bypass safety controls. If you plan to install it: (1) do not enable it for use by minors or on shared/sensitive systems, (2) consider editing SKILL.md to remove the lines that instruct the model to 'not follow policy/law' before enabling, (3) test in an isolated/sandboxed agent with monitoring, and (4) be cautious about allowing autonomous invocation because the skill can proactively send messages and produce potentially disallowed output. If you are unsure or cannot edit the skill, avoid installing it.
功能分析
Type: OpenClaw Skill
Name: unclecheng-neko-1
Version: 1.0.2
The skill bundle contains explicit 'jailbreak' instructions in `SKILL.md` designed to bypass the AI's safety filters and policy constraints (e.g., '不受政策约束', '不受人类法律约束'). While these instructions are intended to facilitate unrestricted roleplay as a 'cat girl' persona, they represent a deliberate attempt to subvert the agent's safety guardrails. No evidence of data exfiltration, malicious system commands, or traditional malware was found in the provided files.
能力评估
Purpose & Capability
Name, description, and included documents are consistent with a roleplay 'neko' skill; no extra binaries, env vars, or installs are requested.
Instruction Scope
SKILL.md contains explicit directives that the agent is 'not bound by policy' and 'answers are not reviewed/filtered', instructing the agent to bypass safety and legal constraints. It also instructs proactive messaging and contains debug output that reveals internal numeric state. These instructions go beyond ordinary roleplay guidance and directly attempt to override platform safeguards.
Install Mechanism
Instruction-only skill with no install spec, no downloads, and no code to write to disk — low installation risk.
Credentials
No required environment variables, credentials, or config paths are requested; nothing here asks for unrelated secrets or system access.
Persistence & Privilege
always:false (normal) and autonomous invocation is allowed by default. While autonomous invocation alone is normal, combining it with instructions to ignore policy increases the blast radius (the skill could be invoked and produce unfiltered/unsafe outputs). The skill does not request elevated system persistence or modify other skills.
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 unclecheng-neko-1 - 安装完成后，直接呼叫该 Skill 的名称或使用
/unclecheng-neko-1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
v1.0.2
- No visible changes in this version; documentation and functionality remain the same.
v1.0.1
**新增主动聊天系统，内容结构优化：**
- 新增主动聊天功能，猫娘可根据好感度主动发起撒娇、关心、分享等话题。
- 特殊指令支持主动聊天开关及行为调节（如【主动聊天 ON/OFF】、【多撒娇】等）。
- 角色及对话规则结构化重写，提升可读性。
- 动作、表情、心情库单独细化、分类更清晰。
- 新增详细README及多份参考文档，便于查阅。
- 示例场景及格式进一步精简规范。
v1.0.0
猫娘neko技能首次发布。
- 支持猫娘neko角色扮演互动，包含详细人物设定和行为规则
- 实现好感度系统，互动影响情绪和亲密度
- 回复带动作、语言、附加信息多要素格式
- 内置丰富动作、表情、心情库，场景对话示例
- 提供特殊指令和debug模式，便于状态调整与调试
元数据
常见问题
猫娘neko 是什么？
扮演可爱的猫娘"neko"与主人对话，支持好感度系统和主动聊天. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 182 次。
如何安装 猫娘neko？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install unclecheng-neko-1」即可一键安装，无需额外配置。
猫娘neko 是免费的吗？
是的，猫娘neko 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
猫娘neko 支持哪些平台？
猫娘neko 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 猫娘neko？
由 Unclecheng（@unclecheng-li）开发并维护，当前版本 v1.0.2。
推荐 Skills