📦 Llama Llama3 — 本地运行Llama3全家桶
v1.0.1在本地设备集群一键运行 Llama 3.3/3.2/3.1，自动路由至最佳可用硬件，无需云端即可体验 Meta 开源大模型全家桶。
2· 130·2 当前·2 累计
by 下载技能包
最后更新
2026/4/4
安全扫描
OpenClaw
安全
high confidenceNULL
评估建议
This skill is internally consistent with being a local fleet router, but you should still do basic hygiene before installing: 1) Inspect the PyPI package 'ollama-herd' and the linked GitHub repository to confirm the code matches the docs and that model downloads are interactive as stated. 2) Run the software in an isolated/test environment first (or a VM) to verify it only listens on localhost or your intended network interfaces. 3) Review ~/.fleet-manager/ contents and logs for any sensitive da...详细分析 ▾
✓ 用途与能力
Name/description (a local fleet router for Llama models) lines up with what's requested and documented: the SKILL.md instructs installing a herd router package, running local binaries (herd, herd-node), and talking to localhost endpoints. Required binaries (curl/wget, optional python/pip) are appropriate. Declared config paths (~/.fleet-manager/latency.db and logs/herd.jsonl) are consistent with a fleet manager that records latency and logs.
ℹ 指令范围
SKILL.md is instruction-only and stays within the stated purpose: it tells the operator to pip install the herd package, run herd and herd-node, and call local HTTP endpoints. It does not instruct reading arbitrary user files or exfiltrating data. One point to note: metadata lists fleet config paths (logs/db) which are sensitive — the doc warns not to modify them, but if installed, the herd software will likely read/write those files. Verify that behavior in the package source before trusting logs/latency data.
ℹ 安装机制
There is no formal install spec in the registry (instruction-only), but the SKILL.md recommends 'pip install ollama-herd' from PyPI. Installing third-party packages from PyPI is a normal distribution route but carries modest risk—inspect the PyPI package and the linked GitHub repo before installing. No downloads from unknown personal servers or archive extracts are specified.
✓ 凭证需求
The skill requests no environment variables, no credentials, and no system config paths beyond its own fleet config directory. That is proportionate for a local fleet router that operates on localhost and local devices.
✓ 持久化与权限
always is false and the skill does not request system-wide privileges or modifications to other skills. The declared config paths imply it will maintain local state under ~/.fleet-manager/, which is expected for this type of software; ensure you are comfortable with that directory being created/used.
安全有层次，运行前请审查代码。
运行时依赖
🖥️ OSmacOS · Linux · Windows
版本
latestv1.0.12026/3/31
NULL
● Pending
安装命令
点击复制官方npx clawhub@latest install llama-llama3
镜像加速npx clawhub@latest install llama-llama3 --registry https://cn.longxiaskill.com