Token导航 LogoToken导航

Agent Skill · 图像处理

lumina-image(亮度图像)

用于辅助图像生成、图片编辑、视觉素材处理或图像模型工作流。它适合让 Agent 根据文本生成图片、处理背景、整理视觉提示词或调用相关图像工具。使用时需要确认输入图片、版权来源、输出格式和模型限制;涉及人物、品牌、商品或公开展示素材时,应额外核对授权、真实性和内容合规边界。

lumina-image 是图像处理类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为需要联网;许可证信息还不明确,第三方扫描已通过,按安装量在图像处理类的 1,562 个 Skill 里排第 1,451。

总安装 279周安装 12GitHub Stars 公开资料未说明下载 98核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
unknown

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:lumina-image(亮度图像)
来源仓库:https://github.com/liulixiang1988/agent-skills
仓库路径:skills/lumina-image
安装命令:
npx skills add https://github.com/liulixiang1988/agent-skills --skill lumina-image
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/liulixiang1988/agent-skills --skill lumina-image

这个 Skill 是做什么的

图像处理类 Skill 处理生成、裁剪、压缩和批量转换图片的任务。

lumina-image 用于辅助图像生成、图片编辑和视觉素材处理,适合在 Codex、Claude、Cursor、Gemini CLI 中调用图像工具。适用于文本生成图片、背景处理或视觉提示词整理等图像相关工作。通过 npx skills add 命令安装,需结合项目实际使用场景验证功能。

使用时需确认输入图片版权、输出格式限制,涉及人物或品牌素材时应核对授权与合规性。适用宿主包括 Codex、Claude、Cursor、Gemini CLI,接入前应确认版本、权限和运行环境要求。

来源仓库内路径:skills/lumina-image

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/liulixiang1988/agent-skills --skill lumina-ima…;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它需要需要联网,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

lumina-image 怎么安装?

在终端执行 npx skills add https://github.com/liulixiang1988/agent-skills --skill lumina-image。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

lumina-image 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

lumina-image 需要什么权限?

权限级别是需要联网。该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

lumina-image 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

lumina-image 用的人多吗?

公开数据是累计安装 279 次,近一周安装 12 次,下载 98 次。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

lumina-image 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 第三方扫描已通过,常见的高风险行为没有命中。

先想清楚的地方

  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理图像处理类任务、又不想每次都手写提示词的人。

不太适合

  • 需要明确授权才能用到公司项目里的场景。
  • 需要厂商级技术支持和 SLA 的生产系统。

图像处理类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
ai-image-generation

AI 图像生成 · Codex、Claude

4,365,552382调用外部服务未明确有提醒
openai-image-gen

OpenAI 图像生成 · OpenClaw

1,906,88432需要联网MIT-0有提醒
image

图像 · OpenClaw

525,99926需要联网MIT-0通过
baoyu-cover-image

宝玉封面图 · Codex、Claude

474,24016,779只读MIT有提醒
baoyu-image-gen

宝玉图像生成器 · Codex、Claude

468,00716,778接触敏感数据MIT通过
lumina-image

亮度图像 · 本页

279—需要联网未明确通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

Lumina Image Build Commands

You build Lumina-specific container images by sourcing and calling pre-built script functions. Do NOT manually replicate the steps in the scripts — just source and call the function directly.

Build Proxy API Image (Windows Only)

When the user asks to build the proxy API image (e.g., "build proxy api image", "build lumina proxy", "build proxy api"):

if (-not $Env:MS_PATH) { $Env:MS_PATH = Get-Location }
. "<skill-path>/scripts/build-proxy.ps1"; lumina_build_proxy_api_image

Build SandboxControlPlane API Image (Windows Only)

When the user asks to build the SandboxControlPlane API image (e.g., "build scp api image", "build sandboxcontrolplane image", "build scp image", "build sandbox control plane api"):

if (-not $Env:MS_PATH) { $Env:MS_PATH = Get-Location }
. "<skill-path>/scripts/build-proxy.ps1"; lumina_build_scp_api_image

Build Sandbox Agent Image

When the user asks to build the sandbox agent image (e.g., "build sandbox agent", "build sandbox agent image", "build sandbox image"):

Windows (PowerShell):

if (-not $Env:MS_PATH) { $Env:MS_PATH = Get-Location }
. "<skill-path>/scripts/build-proxy.ps1"; sandbox_build_agent_image

macOS/Linux (bash/zsh):

if [ -z "${MS_PATH:-}" ]; then export MS_PATH="$(pwd)"; fi
. "<skill-path>/scripts/build-sandbox-agent.sh"; sandbox_build_agent_image

Replace Kubernetes Deployment Image

When the user asks to replace or update an existing Lumina deployment image in Kubernetes after building an image (for example, replacing the SandboxControlPlane deployment image in AKS), do not change the deployment immediately.

You must first ask the user to confirm the target values for all of the following:

  • Kube context: Ask the user which kubectx / Kubernetes context to use.
  • Namespace: Ask the user which namespace to use.
  • Deployment: Ask the user which deployment to update.
  • Image: Confirm the exact target image tag that will be applied.

After collecting those values, explicitly ask the user whether to replace the existing deployment image. Only perform the kubectl set image / rollout step after the user confirms.

Recommended execution flow:

  1. Ask the user to choose the kube context.
  2. Ask the user to choose the namespace.
  3. Ask the user to choose the deployment.
  4. Show the exact image that will be used.
  5. Ask for confirmation to replace the existing deployment image.
  6. Only after confirmation, update the deployment image and verify rollout status.

Recommended question template:

I can replace the existing deployment image, but I need you to confirm the target first.

Please choose:
1. kube context
2. namespace
3. deployment

Then I will show the exact image tag and ask for final confirmation before applying the change.

Recommended confirmation template:

I am about to replace the image for deployment <deployment> in namespace <namespace> on context <kube-context>.

Target image:
<image>

Do you want me to replace the existing deployment image now?

Recommended execution commands after confirmation:

kubectl --context <kube-context> -n <namespace> get deployment <deployment> -o json
kubectl --context <kube-context> -n <namespace> set image deployment/<deployment> <container-name>=<image>
kubectl --context <kube-context> -n <namespace> rollout status deployment/<deployment> --timeout=10m
kubectl --context <kube-context> -n <namespace> get pods -l app=<deployment> -o wide

Notes:

  • Read the deployment first to identify the correct container name before running kubectl set image.
  • Do not assume the container name is the same as the deployment name.
  • After rollout, verify that the new pod is running and that the deployment template now points to the requested image.

If the deployment has multiple containers:

  • List all container names and their current images before making any change.
  • Show the list to the user and ask which container should be updated.
  • Do not update multiple containers unless the user explicitly asks for that.
  • Repeat the final confirmation step with the selected container name included.

Recommended multi-container inspection command:

kubectl --context <kube-context> -n <namespace> get deployment <deployment> -o jsonpath="{range .spec.template.spec.containers[*]}{.name}{'|'}{.image}{'\n'}{end}"

Recommended multi-container question template:

This deployment has multiple containers. Please choose which container to update:

<container-name-1> | <current-image-1>
<container-name-2> | <current-image-2>
<container-name-3> | <current-image-3>

Recommended multi-container confirmation template:

I am about to replace the image for container <container-name> in deployment <deployment> within namespace <namespace> on context <kube-context>.

Current image:
<current-image>

Target image:
<target-image>

Do you want me to replace this container image now?

Prerequisites

  • MS_PATH: Must point to the CopilotLumina root directory. If not set, the commands above auto-detect it from the current working directory.
  • ACR Login: If a docker push fails with an authentication error, run az acr login -n luminaacrdev and retry.
  • <skill-path>: Replace with the actual path to the skill directory containing the scripts folder.
  • Kubernetes updates: Before replacing a deployment image, ask the user to choose the kube context, namespace, and deployment, then ask for explicit confirmation before applying the change.

Behavior

  • lumina_build_proxy_api_image remains Windows/PowerShell only.
  • lumina_build_scp_api_image is Windows/PowerShell only.
  • sandbox_build_agent_image now supports both Windows (PowerShell) and macOS/Linux (bash/zsh).
  • All commands build the image, push to ACR, and print the image tag.
  • For Kubernetes deployment image replacement, the skill must gather the kube context, namespace, and deployment from the user via questions, and must not replace an existing deployment image without explicit user confirmation.

平台分布

Codex

32.17%
按下载量换算32

Claude

30.55%
按下载量换算30

Cursor

19.48%
按下载量换算19

Gemini CLI

9.39%
按下载量换算9

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

需要联网

该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills