← 返回 Skills 市场

Jimeng AI Image Generation

Name: Jimeng AI Image Generation
Author: ken0521

作者多动朕 · GitHub ↗ · v1.0.0 · MIT-0

cross-platform ✓ 安全检测通过

1479

总下载

当前安装

版本数

在 OpenClaw 中安装

/install jimeng-image-gen

功能描述

即梦 AI 图片生成技能（火山引擎图片生成 4.0）。当用户想要 AI 生成图片、文生图、图生图、字体设计、海报制作时使用。支持场景： - "帮我生成一张图片：..." - "用即梦画一张 16:9 的科技感壁纸" - "字体设计：新年快乐，红色背景" - "把这张图的背景换成星空" - "生成一组表情包" -...

使用说明 (SKILL.md)

即梦 AI 图片生成（4.0）

基于火山引擎即梦 AI 图片生成 4.0 异步接口，纯 Node.js 内置模块，零依赖。

快速使用

# 文生图（2K 自动比例）
node {baseDir}/scripts/jimeng.mjs generate --prompt "赛博朋克城市夜景，霓虹灯，雨天"

# 指定比例
node {baseDir}/scripts/jimeng.mjs generate --prompt "山水画，中国风" --ratio 16:9

# 4K 高清
node {baseDir}/scripts/jimeng.mjs generate --prompt "星空下的雪山" --ratio 16:9-4k --save ./output.png

# 字体设计模式
node {baseDir}/scripts/jimeng.mjs generate --text "618大促" --color 橙色 --illustration "购物车,礼盒,星星" --ratio 4:3

# 图生图（换背景）
node {baseDir}/scripts/jimeng.mjs generate --prompt "把背景换成星空" --image-url "https://..." --scale 0.7

# 强制单图（快速省钱）
node {baseDir}/scripts/jimeng.mjs generate --prompt "一只猫" --force-single --save ./cat.png

# 查看比例预设
node {baseDir}/scripts/jimeng.mjs ratios

参数说明

提示词（二选一）

参数	说明
`--prompt`	通用文生图提示词，中英文均可，最长 800 字符
`--text`	字体设计模式：图片上显示的文字
`--color`	字体设计模式：背景主色调（默认: 蓝色）
`--illustration`	字体设计模式：配饰元素，逗号分隔

分辨率（三选一，不传默认 2K 自动比例）

参数	说明
`--ratio`	比例预设（见 `ratios` 命令），如 `16:9`、`4:3`、`1:1`、`16:9-4k`
`--width` + `--height`	精确宽高（需同时传）
`--size`	像素面积，如 `4194304`（2K）、`16777216`（4K）

图生图

参数	说明
`--image-url`	参考图 URL，多张用逗号分隔，最多 10 张
`--scale`	文本影响强度 0~1（0=完全参考图，1=完全按提示词，默认 0.5）

其他

参数	说明
`--force-single`	强制只生成 1 张（省钱省时，避免 AI 自动多图计费）
`--save \x3C路径>`	保存图片到本地，多图自动加序号

注意事项

计费：按输出图片张数计费，AI 默认可能输出多张；对费用敏感时加 --force-single
图片 URL 有效期：24 小时，需及时下载
任务超时：最长等待 120 秒，超时后可用 task_id 手动查询
内容审核：提示词含敏感词会被拦截（code 50412/50413），修改提示词后重试

API 详情

见 references/api.md，包含完整参数表、错误码、计费说明。

安全使用建议

This skill appears to be a straightforward Node.js client for Volcengine '即梦' image generation. Before installing/using: 1) Understand that all prompts and any image URLs you supply will be sent to volcengine (visual.volcengineapi.com) and the generated images will be downloaded — do not send private or sensitive images or prompts you do not want sent to an external service. 2) Provide dedicated API keys with the minimum necessary privileges and monitor usage/billing (the service bills per output image). 3) The script will write files to disk when you use --save; run it in a directory you control. 4) Review the included scripts yourself (they are present and human-readable) before running. 5) If you need stronger assurance about provenance, obtain the skill from a known publisher or the official vendor SDK rather than an unknown source.

功能分析

Type: OpenClaw Skill Name: jimeng-image-gen Version: 1.0.0 The skill bundle provides a legitimate interface for the Volcengine Jimeng AI image generation service. The main script (scripts/jimeng.mjs) implements the standard HMAC-SHA256 V4 signing algorithm to communicate with the official endpoint (visual.volcengineapi.com) and uses only built-in Node.js modules. No evidence of data exfiltration, malicious execution, or prompt injection was found; the script's behavior is entirely consistent with its stated purpose of generating and downloading images.

能力评估

✓ Purpose & Capability

Name/description match the implementation: the skill is a Node.js client for Volcengine '即梦' image-generation (jimeng) and correctly requires node and the service Access/Secret keys. Required items (node, JIMENG_ACCESS_KEY, JIMENG_SECRET_KEY) are proportionate and expected.

✓ Instruction Scope

SKILL.md instructs running the included Node script which reads only the declared env vars, sends prompts/optional image URLs to the official visual.volcengineapi.com endpoints, polls for results, and can save returned images to disk. The script makes network requests to the vendor and to image URLs returned by the service (for downloads) — this is expected for the stated purpose.

✓ Install Mechanism

No install spec and the package is implemented as a single Node script using only built-in Node modules (crypto, https, fs, etc.). Nothing is downloaded from arbitrary or obscure URLs and no archive extraction or package manager installs are required.

✓ Credentials

Only two env vars are required (ACCESS and SECRET) which are exactly the credentials needed to sign requests to the service. No unrelated secrets or system credentials are requested.

✓ Persistence & Privilege

The skill does not request always:true, does not modify other skills or system settings, and is user-invocable. It runs on-demand and does not demand persistent elevated privileges.

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 jimeng-image-gen
安装完成后，直接呼叫该 Skill 的名称或使用 /jimeng-image-gen 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

v1.0.0

Initial release: Jimeng AI image generation based on Volcengine 4.0 async API. Supports text-to-image, image-to-image, font design poster mode, up to 4K resolution, AI auto ratio inference. Zero dependencies, pure Node.js built-in modules.

元数据

Slug jimeng-image-gen

版本 1.0.0

许可证 MIT-0

累计安装 3

当前安装数 3

历史版本数 1

常见问题