← 返回 Skills 市场
Qwen Qwen3 5
作者
Twin Geeks
· GitHub ↗
· v1.0.1
· MIT-0
155
总下载
1
收藏
2
当前安装
2
版本数
在 OpenClaw 中安装
/install qwen-qwen3-5
功能描述
Qwen 3.5 by Alibaba — run Qwen 3.5 (the latest and most capable Qwen model) across your local device fleet. Qwen 3.5 rivals GPT-4o and Claude 3.5 on reasonin...
安全使用建议
This skill is internally consistent with its purpose, but before installing: (1) review the referenced GitHub repo (https://github.com/geeks-accelerator/ollama-herd) and the PyPI package (ollama-herd) to verify maintainers and code, (2) be aware that `pip install` will execute third-party code—run it in an isolated environment or VM if you have doubts, (3) inspect ~/.fleet-manager files if they already exist (they may contain logs/state), (4) note that starting the router opens a local network service on port 11435—ensure it is bound to localhost and protected by firewall rules if you don’t want remote access, and (5) confirm model downloads and deletions actually require explicit confirmation as claimed. Because there is no bundled code in the skill bundle itself, the main risk comes from the external package and runtime software it instructs you to install.
能力评估
Purpose & Capability
Name/description (run Qwen 3.5 on a local fleet) align with what's requested: curl/wget for pulls and optional python/pip for the referenced Ollama Herd tooling. Declared config paths (~/.fleet-manager/...) are plausible for a fleet router.
Instruction Scope
SKILL.md instructs installing and running the ollama-herd tooling (pip install ollama-herd; herd; herd-node) and calling a local HTTP endpoint (localhost:11435). It does not ask the agent to read unrelated system files or external secrets, but it does reference and rely on ~/.fleet-manager files and local network endpoints (dashboard/API) which could contain logs or state.
Install Mechanism
No registry install spec is provided; the instructions tell the user to run `pip install ollama-herd` (PyPI). Installing a third-party package is a normal vector for this functionality but carries moderate risk because arbitrary code will be downloaded and executed from PyPI; the skill itself does not include code to audit.
Credentials
The skill requests no credentials or environment variables. The only nonstandard things are two config paths under ~/.fleet-manager (latency.db and logs/herd.jsonl) which are consistent with a fleet manager but may contain sensitive logs or metadata; their presence is explainable by the stated purpose.
Persistence & Privilege
always:false and normal model-invocation settings. The skill does not request persistent platform-wide privileges and does not indicate modification of other skills or system-wide agent settings.
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 qwen-qwen3-5 - 安装完成后，直接呼叫该 Skill 的名称或使用
/qwen-qwen3-5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
v1.0.1
Cross-platform support: macOS, Linux, and Windows. Updated OS metadata, descriptions, and hardware recommendations.
v1.0.0
Initial release of Qwen 3.5 skill for Ollama Herd.
- Adds support for running Alibaba's Qwen 3.5 (72B, 32B, 14B, 7B), Qwen3-Coder, and Qwen3-ASR models across local device fleets.
- Fleet-aware routing to the best machine; no cloud costs or automatic model downloads.
- Provides API compatibility with OpenAI SDK and Ollama API.
- Includes documentation for setup, usage, and hardware recommendations.
- Strong privacy, no rate limits, and user-controlled model management.
- Optional use of Python tooling; supports macOS and Linux.
元数据
常见问题
Qwen Qwen3 5 是什么？
Qwen 3.5 by Alibaba — run Qwen 3.5 (the latest and most capable Qwen model) across your local device fleet. Qwen 3.5 rivals GPT-4o and Claude 3.5 on reasonin... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 155 次。
如何安装 Qwen Qwen3 5？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install qwen-qwen3-5」即可一键安装，无需额外配置。
Qwen Qwen3 5 是免费的吗？
是的，Qwen Qwen3 5 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
Qwen Qwen3 5 支持哪些平台？
Qwen Qwen3 5 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（darwin, linux, windows）。
谁开发了 Qwen Qwen3 5？
由 Twin Geeks（@twinsgeeks）开发并维护，当前版本 v1.0.1。
推荐 Skills