← 返回 Skills 市场
132
总下载
0
收藏
0
当前安装
2
版本数
在 OpenClaw 中安装
/install tunelo
功能描述
Expose local services and files to the internet through a public HTTPS URL. Designed for AI agents — when you need to let a user preview files remotely, shar...
安全使用建议
Before installing or running this skill, consider the following: (1) do not run `curl | sh` from an unknown domain — request source code, a release page, and a cryptographic checksum (or install from a trusted package repository); (2) understand that exposing a directory or service publishes whatever is served — double-check paths to avoid leaking secrets, private keys, config files, or other sensitive data; (3) prefer using the `--local` option or a self-hosted relay (`--relay`) if you need testing without using a third-party relay; (4) if you must try it, run the installer and tunelo binary inside a disposable VM or container, or audit the install script first; (5) be cautious about allowing the agent to autonomously execute these install/run commands — require explicit user confirmation each time. If the publisher can supply a verifiable GitHub release, binary checksums, or an auditable install package, that would raise confidence and could change this assessment.
功能分析
Type: OpenClaw Skill
Name: tunelo
Version: 0.2.0
The skill facilitates exposing local services and files to the public internet via a third-party relay (tunelo.net), which is a high-risk capability that can be easily abused for data exfiltration. It explicitly instructs the AI agent to use a 'curl | sh' installation method (https://tunelo.net/install.sh), a dangerous pattern for unverified remote code execution. While the stated purpose is for sharing and previews, the tool provides a functional backdoor into the local environment if the agent is prompted to serve sensitive directories like the user's home folder.
能力评估
Purpose & Capability
The name, description, and commands all line up: this skill's goal is to expose local services/files via a public HTTPS URL and the SKILL.md explains commands to do so. Required capabilities (network relay, ability to serve files) are consistent with the stated purpose.
Instruction Scope
The SKILL.md explicitly instructs the agent to run an installer via `curl -fsSL https://tunelo.net/install.sh | sh` and then to expose arbitrary local paths (e.g., `tunelo serve /path/to/files`). Running an opaque remote install script and exposing arbitrary local directories are both high-risk actions: the install script can execute arbitrary code on the host, and the tunneling commands can publish sensitive files if used incorrectly. The instructions do not provide a checksum, source repository, or details to verify the installer.
Install Mechanism
There is no install spec other than a curl|sh from tunelo.net — a single-file download-and-execute from an unverified domain. Per the scanning rules, this is a high-risk install pattern (arbitrary code and binaries will be written to disk). The domain is not a known, verifiable release host (e.g., GitHub releases, official distro repos) and no integrity verification is provided.
Credentials
The skill does not request environment variables, credentials, or config paths. That is proportional: exposing local services does not inherently require additional external credentials. However, the default behavior routes traffic through a third-party relay (tunelo.net), which implicitly gives that operator visibility into the endpoints you expose — a privacy/operational concern even though not expressed as credentials.
Persistence & Privilege
The install writes a persistent binary (claimed at /usr/local/bin/tunelo). Installing a binary is a persistent change and may require elevated permissions; this is not inherently malicious but is a meaningful system modification. The skill is not set to always:true and does not request platform-wide privileges otherwise.
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 tunelo - 安装完成后,直接呼叫该 Skill 的名称或使用
/tunelo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
v0.2.0
Rewritten for AI agents — trigger phrases, typical workflows, when to use/not use
v0.1.0
Initial release — expose localhost to the internet via QUIC tunnel
元数据
常见问题
tunelo 是什么?
Expose local services and files to the internet through a public HTTPS URL. Designed for AI agents — when you need to let a user preview files remotely, shar... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 132 次。
如何安装 tunelo?
在 OpenClaw 或 Claude Code 对话框中运行命令「/install tunelo」即可一键安装,无需额外配置。
tunelo 是免费的吗?
是的,tunelo 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。
tunelo 支持哪些平台?
tunelo 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 tunelo?
由 claw-bot(@claw-bot)开发并维护,当前版本 v0.2.0。
推荐 Skills