aliyun-image

阿里云百炼图像生成、编辑与翻译。文生图：根据文本生成图像，支持复杂文字渲染。图像编辑：单图编辑、多图融合、风格迁移、物体增删。图像翻译：翻译图像中的文字，保留原始排版，支持11种源语言和14种目标语言。触发词：生成图片、AI作画、文生图、图像编辑、修图、换背景、风格迁移、多图融合、图像翻译、图片翻译。模型：qwen-image-plus(默认)、qwen-image-max、qwen-image-edit-plus(默认)、qwen-image-edit-max、qwen-mt-image。

Security Scan

The skill appears to be an Aliyun (DashScope) image-generation/translation client and its code matches that purpose, but the published metadata omits the required API credential and homepage/source provenance is incomplete — this inconsistency and the fact the client can upload local files merit caution.

Purpose & Capability

Name/description, docs, and client code all target Aliyun DashScope image APIs (generation, edit, translate) and use endpoints under dashscope.aliyuncs.com — that part is coherent. However, registry metadata lists no required environment variables or primary credential, while SKILL.md/README/scripts clearly require DASHSCOPE_API_KEY. The skill's Source/Homepage are marked unknown/none despite repository links inside SKILL.md/README, which is an inconsistency in provenance information.

Instruction Scope

Runtime instructions and code show normal API usage: POST to DashScope endpoints, async task polling, and optional local-file-to-base64 encoding for uploads. The instructions do not ask the agent to read unrelated system files or secrets. Important behavior: the client will read local image files (if given) and encode+upload them, and will download generated images, which is expected for image-edit/translate but carries privacy implications.

Install Mechanism

No install spec (instruction-only) which minimizes installer risk. There is nevertheless a client.py bundled in the package; no install steps are provided to add system binaries. This is consistent with an instruction-only skill but it's worth noting code files are present even though no installation is declared.

Credentials

The skill requires a single API key (DASHSCOPE_API_KEY) according to SKILL.md/README/scripts, which is proportionate to the described cloud API usage. However the declared registry requirements list no env vars or primary credential — a clear mismatch. The client only uses that single key (no unrelated credentials), so the requested privilege is limited but the omission in metadata is a red flag.

Persistence & Privilege

always:false and default invocation settings; no persistent or elevated privileges are requested. The skill does not modify other skills or system-wide configs. Autonomous invocation is allowed (platform default) but not combined with other high-risk flags.

What to consider before installing

This package appears to implement Aliyun DashScope image APIs and will work if you provide a DASHSCOPE_API_KEY, but the public metadata omitted that required environment variable and the homepage/source fields are incomplete. Before installing: 1) Verify the upstream repository (the SKILL.md points to a GitHub URL) and inspect its commits and issues to confirm legitimacy. 2) Supply a least-privilege API key (create a key with only the needed AIGC scopes) and avoid using broader account keys. 3) Remember local files you pass to the client may be uploaded to Aliyun — don't send sensitive images. 4) Test in an isolated environment first and monitor network requests (ensure calls go to dashscope.aliyuncs.com). If the missing DASHSCOPE_API_KEY declaration or absent homepage concerns you, ask the publisher to correct the metadata before proceeding.

阿里云百炼图像生成与编辑

🔄 同步更新：本项目在 GitHub 和 ClawHub 同步发布。

快速开始

前提条件

# 配置 API Key
export DASHSCOPE_API_KEY="sk-xxxx"

文生图

import os, requests

resp = requests.post(
    "https://dashscope.aliyuncs.com/api/v1/services/aigc/multimodal-generation/generation",
    headers={"Authorization": f"Bearer {os.getenv('DASHSCOPE_API_KEY')}"},
    json={
        "model": "qwen-image-plus",
        "input": {"messages": [{"role": "user", "content": [{"text": "一只橘猫在阳光下打盹"}]}]},
        "parameters": {"prompt_extend": True, "watermark": False}
    }
)
print(resp.json()["output"]["choices"][0]["message"]["content"][0]["image"])

图像编辑

resp = requests.post(
    "https://dashscope.aliyuncs.com/api/v1/services/aigc/multimodal-generation/generation",
    headers={"Authorization": f"Bearer {os.getenv('DASHSCOPE_API_KEY')}"},
    json={
        "model": "qwen-image-edit-plus",
        "input": {"messages": [{"role": "user", "content": [
            {"image": "https://example.com/photo.jpg"},
            {"text": "把背景换成星空"}
        ]}]},
        "parameters": {"prompt_extend": True, "watermark": False}
    }
)
print(resp.json()["output"]["choices"][0]["message"]["content"][0]["image"])

图像翻译

import os, requests, time

# 1. 创建翻译任务
resp = requests.post(
    "https://dashscope.aliyuncs.com/api/v1/services/aigc/image2image/image-synthesis",
    headers={
        "Authorization": f"Bearer {os.getenv('DASHSCOPE_API_KEY')}",
        "X-DashScope-Async": "enable"
    },
    json={
        "model": "qwen-mt-image",
        "input": {
            "image_url": "https://example.com/english-poster.jpg",
            "source_lang": "en",
            "target_lang": "zh"
        }
    }
)
task_id = resp.json()["output"]["task_id"]

# 2. 轮询获取结果
while True:
    time.sleep(3)
    result = requests.get(
        f"https://dashscope.aliyuncs.com/api/v1/tasks/{task_id}",
        headers={"Authorization": f"Bearer {os.getenv('DASHSCOPE_API_KEY')}"}
    ).json()
    if result["output"]["task_status"] == "SUCCEEDED":
        print(result["output"]["image_url"])
        break

---

默认配置

场景	默认模型	高质量模型
文生图	qwen-image-plus	qwen-image-max
图像编辑	qwen-image-edit-plus	qwen-image-edit-max
图像翻译	qwen-mt-image	-

规则：默认使用 Plus 系列，仅当用户明确要求"最好的"、"最高质量"时使用 Max 系列。

默认参数

{
  "negative_prompt": "低分辨率，低画质，肢体畸形，手指畸形，画面过饱和，蜡像感",
  "prompt_extend": true,
  "watermark": false
}

---

详细参考

按需加载以下文档：

文档	用途
references/text-to-image.md	文生图完整API参考：模型列表、分辨率、所有参数
references/image-edit.md	图像编辑完整API参考：单图编辑、多图融合、输入要求
references/image-translate.md	图像翻译完整API参考：支持语言、异步调用、术语定义

---

工具脚本

脚本	用途
scripts/client.py	封装好的API客户端，支持文生图、图像编辑和图像翻译

使用脚本：

# 文生图
python scripts/client.py generate "一只橘猫在阳光下打盹" --size 1920*1080

# 图像编辑
python scripts/client.py edit "https://example.com/photo.jpg" "把背景换成星空" -n 2

# 图像翻译
python scripts/client.py translate "https://example.com/english.jpg" --source en --target zh

---

关键注意事项

1. 图像URL有效期：生成的图像URL仅保留 24小时，请及时下载
2. 地域：API Key 和请求地址必须属于同一地域（北京/新加坡）
3. 费用：按成功生成的图像张数计费

---

常见错误

错误码	原因	解决方案
InvalidApiKey	API Key无效	检查环境变量配置
InvalidParameter	参数错误	检查请求格式
DataInspectionFailed	内容审核失败	修改输入内容