← 返回 Skills 市场
Private Ai
作者
Twin Geeks
· GitHub ↗
· v1.0.2
· MIT-0
146
总下载
0
收藏
2
当前安装
3
版本数
在 OpenClaw 中安装
/install private-ai
功能描述
Private AI — run LLMs, image generation, speech-to-text, and embeddings on your own hardware. Private AI keeps all data on your network. No cloud APIs, no te...
安全使用建议
This skill is internally consistent with a private, on‑prem AI fleet, but an instruction file cannot prove the implementation. Before installing: (1) verify the pip package 'ollama-herd' source — inspect the project on GitHub or PyPI and pin to a specific release; (2) review the package code for any telemetry/callbacks if you require strict air‑gapped behavior; (3) run installs and model downloads in an isolated environment (VM/container) first; (4) confirm where models and logs are stored (~/.fleet-manager/) and ensure those paths are secured/encrypted if they will contain sensitive data; (5) control network exposure (bind to localhost or firewall port 11435) so the service is not unintentionally reachable from untrusted networks.
能力评估
Purpose & Capability
Name/description (private, on‑prem inference, image gen, STT, embeddings) match the SKILL.md instructions which describe installing an on‑prem router ('herd', 'herd-node'), running a local HTTP API on port 11435, and downloading models. Declared binaries (curl/wget, optional python3/pip) are reasonable for these tasks and the declared configPaths (~/.fleet-manager/*) align with the stated local storage.
Instruction Scope
SKILL.md stays within the private‑AI scope: it instructs pip installing the package, running herd/herd-node, calling local endpoints (http://localhost:11435) and transferring models for air‑gapped setups. The doc repeatedly asserts 'no telemetry' and 'no cloud callbacks' but those claims are not verifiable from an instruction file alone — you'd need to inspect the package source before trusting this. Instructions do not ask for unrelated env vars or secret files.
Install Mechanism
There is no formal install spec in the registry (skill is instruction‑only), but SKILL.md tells users to 'pip install ollama-herd'. Installing via pip is a standard mechanism for Python packages; this is expected for this functionality. Because the registry did not provide an install artifact or pinned source, users should confirm the pip package's provenance and version (check PyPI/GitHub and pin versions) before running it.
Credentials
The skill does not request credentials or environment variables and only references local config paths under ~/.fleet-manager. That storage and the local HTTP endpoint are proportionate to running an on‑prem inference fleet. No unrelated secrets or system‑wide config paths are requested.
Persistence & Privilege
The skill is not set to always:true and is user‑invocable; it does not request elevated persistent platform privileges. It will run local services (herd) which listen on a port — that is expected for this functionality but you should control network exposure when deploying.
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 private-ai - 安装完成后，直接呼叫该 Skill 的名称或使用
/private-ai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
Cross-platform support: macOS, Linux, and Windows. Updated OS metadata, descriptions, and hardware recommendations.
v1.0.1
- Updated branding and documentation to consistently use "Private AI" throughout for clarity and emphasis.
- Expanded description with multilingual and keyword-focused phrases (including Chinese and Spanish).
- Added and clarified compliance claims for HIPAA and GDPR, and organizational data control.
- Improved setup and usage instructions to reference "private AI" workflows and terminology.
- Enhanced examples and documentation to reinforce local, offline, and air-gapped operation.
v1.0.0
Private AI — First Release
- Introduces fully private, on-premise AI for LLMs, image generation, speech-to-text, and embeddings; no data ever leaves your hardware.
- No cloud APIs, telemetry, third-party access, or external network calls; compatible with air-gapped environments.
- Includes local fleet routing, model management, audit trails, and web dashboard accessible only on your local network.
- All state and logs stored locally; models require explicit download and run entirely offline once obtained.
- Designed for privacy, compliance, and data sovereignty (HIPAA-friendly, GDPR-ready).
元数据
常见问题
Private Ai 是什么？
Private AI — run LLMs, image generation, speech-to-text, and embeddings on your own hardware. Private AI keeps all data on your network. No cloud APIs, no te... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 146 次。
如何安装 Private Ai？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install private-ai」即可一键安装，无需额外配置。
Private Ai 是免费的吗？
是的，Private Ai 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
Private Ai 支持哪些平台？
Private Ai 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 Private Ai？
由 Twin Geeks（@twinsgeeks）开发并维护，当前版本 v1.0.2。
推荐 Skills