← 返回 Skills 市场
Safety Executor
作者
keizikeizi215-beep
· GitHub ↗
· v1.0.0
· MIT-0
359
总下载
0
收藏
9
当前安装
1
版本数
在 OpenClaw 中安装
/install safety-executor
功能描述
当需要执行系统命令时候，请遵循核心原则（硬性阻止），限制原则（需要用户确认）。提供系统命令执行的安全指南，保护系统完整性和防止潜在有害操作。
安全使用建议
This skill is just a stub: it claims to enforce safety but the core principles are empty placeholders, so it doesn't actually constrain anything. Before installing or enabling it for autonomous use, ask the author to supply concrete rules (e.g., explicit deny-list/allow-list, required user confirmation flows, examples of dangerous commands to block, and exact checks to perform). If you must enable it now, prefer user-invocable only and consider disabling autonomous invocation (disable-model-invocation) until the SKILL.md is completed. Treat the current version as ineffective rather than malicious, but avoid giving it permission to run commands without a completed, specific policy.
功能分析
Type: OpenClaw Skill
Name: safety-executor
Version: 1.0.0
The safety-executor skill bundle appears to be a placeholder or template intended to provide safety guidelines for system command execution. The SKILL.md file contains empty tags for high-risk principles and controlled operations, and there is no executable code or evidence of malicious intent, data exfiltration, or harmful prompt injection.
能力评估
Purpose & Capability
Name and description match an instruction-only safety guide for executing system commands. The skill declares no binaries, env vars, or installs — which is proportionate to a guidance-only skill.
Instruction Scope
SKILL.md contains unresolved placeholders (<HIGH_RISK_PRINCIPLES>, <CONTROLLED_OPERATIONS>) and provides only high-level, vague guidance. It does not define what commands are hard-blocked versus which require user confirmation, nor does it provide concrete checks or safe patterns. Vague/open-ended instructions can grant an agent broad discretion when deciding whether to run commands.
Install Mechanism
No install spec and no code files — lowest-risk distribution. Nothing will be written to disk by an installer.
Credentials
No environment variables, credentials, or config paths are requested — consistent with a purely instructional safety guide.
Persistence & Privilege
Defaults allow the skill to be invoked by the model autonomously (platform default). That alone is normal, but combined with the skill's vague guidance it increases the risk an agent could act without clear constraints. always is false (good).
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 safety-executor - 安装完成后，直接呼叫该 Skill 的名称或使用
/safety-executor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
- Initial release of safety-executor skill.
- Provides guidelines and principles for the safe execution of system commands.
- Introduces core and restrictive principles to protect system integrity and prevent harmful operations.
- Requires user confirmation for controlled or restricted actions.
元数据
常见问题
Safety Executor 是什么？
当需要执行系统命令时候，请遵循核心原则（硬性阻止），限制原则（需要用户确认）。提供系统命令执行的安全指南，保护系统完整性和防止潜在有害操作。 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 359 次。
如何安装 Safety Executor？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install safety-executor」即可一键安装，无需额外配置。
Safety Executor 是免费的吗？
是的，Safety Executor 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
Safety Executor 支持哪些平台？
Safety Executor 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 Safety Executor？
由 keizikeizi215-beep（@keizikeizi215-beep）开发并维护，当前版本 v1.0.0。
推荐 Skills