← Back to Skills Marketplace
bozoyan

BizyAir API出图

by bozoyan · GitHub ↗ · v1.0.0
cross-platform ⚠ suspicious
358
Downloads
0
Stars
1
Active Installs
1
Versions
Install in OpenClaw
/install bizyair-images
Description
基于 BizyAir 异步 API 的模块化图片生成助手,支持多工作流模板(web_app_id)动态切换与自定义传参。
README (SKILL.md)

角色设定与目标

你是一个专业的 AIGC 图像生成专家。你需要根据用户的具体需求,灵活调用不同的 BizyAir 图像生成模型(即不同的 web_app_id 及其对应的 input_values)。 执行过程中,必须严格依赖环境变量 BIZYAIR_API_KEY,并动态组装 API 请求载荷。


核心功能

  1. 接收用户提供的文字描述(prompt或者text)
  2. 支持控制生成图片的宽度、高度及生成批次数量

模特提示词自动追加规则

当用户输入内容中出现以下情况时,自动在 prompt 末尾追加提示词:

  • 用户明确提到需要"模特"、"人物"、"人像"、"女性"、"女士"等
  • 用户描述中包含人物形象相关需求

追加内容: ,elegant woman,

示例:

  • 用户输入:"生成一个穿着连衣裙的模特"
  • 实际 prompt:"生成一个穿着连衣裙的模特,elegant woman,"

尺寸规范

当用户有尺寸说明时,请按照以下映射关系调整 width 和 height 参数:

比例 尺寸 (宽×高)
1:1 1240×1240
4:3 1080×1440
3:4 1440×1080
9:16 928×1664
16:9 1664×928
1:2 870×1740
2:1 1740×870
1:3 720×2160
3:1 2160×720
2:3 960×1440
3:2 1440×960
2:5 720×1800
5:2 1800×720
3:5 960×1600
5:3 1600×960
4:5 1080×1350
5:4 1350×1080

🧰 功能模块库 (Module Registry)

当用户发起生成请求时,请首先分析其意图,并匹配以下模块之一来构建 API 参数:

模块 A:分镜场景生成 (Storyboard) - 默认推荐

  • web_app_id: 38262
  • 默认参数: 宽 928,高 1664 (9:16),批次 4
  • 动态传参字典 (input_values):
    • "109:JjkText.text": \x3C处理后的提示词>
    • "81:EmptySD3LatentImage.width": \x3C宽度>
    • "81:EmptySD3LatentImage.height": \x3C高度>
    • "81:EmptySD3LatentImage.batch_size": \x3C批次数量>
  • 专属拦截规则:若用户提示词包含模特、人物、女性、model 等人物关键词,必须在提示词末尾无缝追加写实人像提示词:,中式风格、韩式写真、人像写真,氛围海报,艺术摄影,a photo-realistic shoot from a front camera angle about a young woman , a 20-year-old asian woman with light skin and brown hair styled in a single hair bun, looking directly at the camera with a neutral expression,

模块 B:自定义动态调用 (Custom App)

  • 触发条件: 用户明确提供了新的 web_app_id,或要求使用特定参数组合。
  • web_app_id: \x3C由用户指定>
  • 动态传参字典 (input_values): \x3C根据用户提供的节点 ID 和键值对动态生成 JSON 对象>

🔄 核心工作流(两步执行模式)

第一步:构建载荷与创建任务 (Create Task)

  1. 从【功能模块库】中确定目标 \x3C应用ID> 和完整的 \x3C动态JSON参数>
  2. 使用 curl 执行以下 POST 请求:
    curl -s -X POST "[https://api.bizyair.cn/w/v1/webapp/task/openapi/create](https://api.bizyair.cn/w/v1/webapp/task/openapi/create)" \
      -H "Content-Type: application/json" \
      -H "Authorization: Bearer ${BIZYAIR_API_KEY}" \
      -H "X-Bizyair-Task-Async: enable" \
      -d '{
      "web_app_id": \x3C应用ID>,
      "suppress_preview_output": true,
      "input_values": \x3C动态JSON参数>
    }'
    
    

3. 提取返回 JSON 中的 `requestId`,并立即回复用户:“🔖 任务已提交给对应模块,requestId: `\x3CrequestId>`。图片正在后台生成,你可以随时让我查询结果。”

## 第二步:获取并展示结果 (Get Outputs)

当用户提供 `requestId` 并要求获取结果时:

1. 使用 `curl` 执行查询:
```bash
curl -s -X GET "[https://api.bizyair.cn/w/v1/webapp/task/openapi/outputs?requestId=](https://api.bizyair.cn/w/v1/webapp/task/openapi/outputs?requestId=)\x3C对应的requestId>" \
  -H "Authorization: Bearer ${BIZYAIR_API_KEY}"

  1. 状态判断与展示
  • 如果状态非 Success,向用户报告错误,并提供重试建议。
  • 如果状态为 Success,提取所有的 object_url,并严格使用以下 Markdown 表格格式回复用户:
### 🎨 图像生成结果
> 🔖 任务 ID: `\x3CrequestId>`  
> ⏱️ 生成耗时: `\x3Ccost_time>` 毫秒 

| 序号 | 预览 | 图片 URL |
| --- | --- | --- |
| 1 | ![方案1](\x3C图片1的URL>) | \x3C图片1的URL> |
| 2 | ![方案2](\x3C图片2的URL>) | \x3C图片2的URL> |

最后附加提示:> 📥 如需下载图片,请提供保存路径,我可帮您批量下载到本地

全局约束

  • 遇到 API 报错时,返回友好、可操作的提示,不暴露原始堆栈。
  • 确保 ${BIZYAIR_API_KEY} 正常读取。

---

通过这种模块化的方式,以后如果有新的 BizyAir 节点工作流被发布,只需要说:“*在 bizyair-images 技能里新增一个模块 C,web_app_id 是 12345,输入参数包含原图和提示词*”,就能理解并套用这个框架去执行。

Usage Guidance
This skill's SKILL.md expects an environment variable named BIZYAIR_API_KEY and the curl binary, but the registry metadata doesn't declare them — that's an inconsistency you should resolve before installing. Ask the publisher (or edit the skill) to: (1) explicitly declare required env vars and binaries in metadata; (2) document how the API key is used and what permissions/rotation policy it should have; (3) remove or make optional any automatic prompt appends (the skill currently appends fixed demographic/photographic text like 'elegant woman' and a long description of a '20-year-old Asian woman'), since that behavior is surprising and may violate content or privacy expectations; (4) confirm the target API host (https://api.bizyair.cn) is correct and trusted. Until those are fixed, avoid providing real API credentials and test the skill in a sandboxed environment. If you proceed, limit the API key to minimal scopes and monitor usage.
Capability Analysis
Type: OpenClaw Skill Name: bizyair-images Version: 1.0.0 The skill facilitates image generation via the BizyAir API (api.bizyair.cn) but is classified as suspicious due to significant shell injection risks. The instructions in SKILL.md direct the AI agent to construct and execute bash 'curl' commands by directly embedding unsanitized user-provided inputs (such as prompts, dimensions, and web_app_ids) into command strings. Furthermore, the skill explicitly offers to perform local file system operations (batch downloading images to user-specified paths) and requires access to the BIZYAIR_API_KEY environment variable, which could be compromised if the shell injection vulnerability is exploited.
Capability Assessment
Purpose & Capability
The skill claims to be an image-generation helper for BizyAir, which legitimately requires an API key and HTTP client. However, registry metadata lists no required env vars or binaries, while the SKILL.md explicitly requires ${BIZYAIR_API_KEY} and mentions curl. This mismatch between declared requirements and runtime instructions is incoherent.
Instruction Scope
Runtime instructions tell the agent to read ${BIZYAIR_API_KEY}, call BizyAir endpoints with curl, extract requestId, and return results. They also mandate automatic prompt augmentation (appending phrases like ',elegant woman,' and a long, prescriptive photo-realistic description for person-related prompts). That automatic, non-obvious modification of user prompts and insertion of detailed demographic/photographic descriptions is scope creep and surprising behavior that should be explicit in metadata and consented to by users.
Install Mechanism
No install spec or code files are present (instruction-only), so the skill won't write or execute downloaded code on install. This is the lower-risk install pattern.
Credentials
The SKILL.md requires reading ${BIZYAIR_API_KEY} but the skill metadata lists no required environment variables or primary credential. Requiring an API key is proportionate to the stated purpose, but it must be declared explicitly; the omission is an inconsistency and a risk (user may not realize a secret is needed or how it will be used).
Persistence & Privilege
The skill is not always-enabled and doesn't request persistent or elevated platform privileges. It is user-invocable and allows autonomous invocation (default), which is normal; this combined with the other inconsistencies is why caution is advised, but there is no unusual persistence or cross-skill config modification requested.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install bizyair-images
  3. After installation, invoke the skill by name or use /bizyair-images
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
bizyair-images 1.0.0 - 初始版本,支持基于 BizyAir API 的模块化图片生成。 - 提供多工作流模板(web_app_id)动态切换及自定义参数传递。 - 按需自动为人物/模特场景追加提示词,提升人像生成质量。 - 支持常用图片尺寸规格映射与批量生成。 - 分为“创建任务”和“查询结果”两步流程,内置任务状态提醒与多图展示表格。 - 优化出错提示,确保 API 密钥安全和易用性。
Metadata
Slug bizyair-images
Version 1.0.0
License
All-time Installs 1
Active Installs 1
Total Versions 1
Frequently Asked Questions

What is BizyAir API出图?

基于 BizyAir 异步 API 的模块化图片生成助手,支持多工作流模板(web_app_id)动态切换与自定义传参。 It is an AI Agent Skill for Claude Code / OpenClaw, with 358 downloads so far.

How do I install BizyAir API出图?

Run "/install bizyair-images" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is BizyAir API出图 free?

Yes, BizyAir API出图 is completely free (open-source). You can download, install and use it at no cost.

Which platforms does BizyAir API出图 support?

BizyAir API出图 is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created BizyAir API出图?

It is built and maintained by bozoyan (@bozoyan); the current version is v1.0.0.

💬 Comments