agentsclimarketplace

Aliyun image gen

Skill yugasun/skills/skills/aliyun-image-gen

Use when generating images with Aliyun Bailian or DashScope Qwen-Image APIs, especially when you need region-aware HTTP calls, async task polling, or downloading temporary image URLs.From its SKILL.md

Install
npx -y skills add yugasun/skills --skill aliyun-image-gen

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

3 things to look at

  • reads credentialsReads from 2 credential sources: `ALIYUN_API_KEY` and 1 more.
  • 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
  • runs commandsInstructs the agent to run 3 commands, including `skills/aliyun-image-gen/scripts/aliyun_image_gen.py generate \ "一张极简风格的科技海报,白底,中心是蓝色折纸鹤" \ --model qwen-image-2.0-pro \ --size 1024*1024 \ --download` and 2 more.

SKILL.md

2.4 KB, 744 tokens by cl100k_base, as published. Nobody here has run it

Aliyun Image Generation

使用阿里云百炼 Qwen-Image 文生图接口生成图片,默认走官方 HTTP API,不依赖 DashScope SDK。

何时使用

  • 需要通过阿里云百炼 API 做文生图。
  • 需要区分北京与新加坡地域 endpoint。
  • 需要提交异步任务、轮询任务状态,或把返回的临时图片 URL 下载到本地。

快速开始

先配置环境变量:

export ALIYUN_API_KEY="sk-xxx"

兼容旧变量名 DASHSCOPE_API_KEY,但优先读取 ALIYUN_API_KEY

同步生成一张图并下载:

skills/aliyun-image-gen/scripts/aliyun_image_gen.py generate \
  "一张极简风格的科技海报,白底,中心是蓝色折纸鹤" \
  --model qwen-image-2.0-pro \
  --size 1024*1024 \
  --download

异步提交并等待完成:

skills/aliyun-image-gen/scripts/aliyun_image_gen.py generate \
  "一张写实风格的咖啡馆照片,暖色夕阳光线" \
  --mode async \
  --model qwen-image-plus \
  --wait \
  --download

按任务 ID 查询异步结果:

skills/aliyun-image-gen/scripts/aliyun_image_gen.py status <task_id> --download

工作流约定

  • 默认地域是 beijing;新加坡需显式传 --region singapore
  • 同步模式默认模型是 qwen-image-2.0-pro,适合大多数文生图调用。
  • 异步模式当前只允许 qwen-image-plusqwen-image,脚本会先做本地校验。
  • 返回图片 URL 默认只有 24 小时有效期;如果要保留结果,立即加 --download

关键参数

  • --negative-prompt: 反向提示词。
  • --size: 输出分辨率,格式如 1024*10241664*928
  • --image-count: 期望输出张数;实际是否支持以模型约束为准。
  • --no-prompt-extend: 关闭模型自动扩写提示词。
  • --watermark: 打开水印,默认关闭。
  • --json: 输出完整 JSON,方便二次处理。

参考资料

<!-- Source references: - https://bailian.console.aliyun.com/cn-beijing?tab=api&productCode=p_efm&switchAgent=13556636#/api/?type=model&url=2975126 -->

What ships with it: 3 files

21.0 KB alongside SKILL.md, 2 of them executable

references/

scripts/

tests/

Keep looking

Skills are one crate of 325,949. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.