← 返回 Skills 市场

豆包AI图像生成

Name: 豆包AI图像生成
Author: fanxi-ju

作者 FANXI-JU · GitHub ↗ · v1.0.0 · MIT-0

cross-platform ⚠ suspicious

183

总下载

当前安装

版本数

在 OpenClaw 中安装

/install doubao-ai-image

功能描述

Generate free AI images via Doubao web interface using automated browser interaction for detailed and styled visual content without API costs.

使用说明 (SKILL.md)

doubao-ai-image

Description

Free AI image generation using Doubao (豆包) web interface through browser automation. Mimics human interaction to generate and download images without API costs.

Usage Scenarios

Free AI image generation for personal or business use
When API-based image generation is not available or too expensive
Automated image creation for reports, presentations, or social media
Integration with other workflows requiring visual content

Workflow

Navigate to Doubao: Open https://www.doubao.com/chat/ in browser
Access Image Generation: Click on "图像生成" or "AI 创作" feature
Input Prompt: Enter detailed image description in the text box
Generate Images: Submit the prompt and wait for AI generation
Select Image: Choose from the generated options (typically 4 variations)
Capture Image: Use browser automation to screenshot the large preview image (ref="preview") to avoid download issues
Save Locally: Store image in /workspace/ai_images/doubao/ directory with timestamp
Deliver: Send image directly to Feishu chat or other specified destination

Key Features

Completely free AI image generation (no API costs)
Human-like browser interaction to avoid bot detection
Supports detailed prompts with style, composition, and quality specifications
Automatic file management and delivery integration
Works with Doubao's Seedream 4.5 model and various aspect ratios/styles

Technical Requirements

Browser automation capability with proper user agent
Element targeting for precise screenshot capture (using ARIA refs)
Image format handling (PNG, JPG, WebP)
Error handling for generation failures or UI changes
Wait mechanisms for AI generation completion
Local file system access for image storage

Best Practices

Use specific, detailed prompts for better results
Include style references when needed (e.g., "realistic", "cartoon", "anime")
Specify aspect ratio if important for the use case
Target the "preview" element (ref="preview") for clean screenshot capture
Verify image quality and composition before delivery
Handle cases where generation fails or produces unsuitable results
Save images to standardized directory structure: /workspace/ai_images/doubao/YYYY-MM-DD/

Limitations

Dependent on Doubao web interface availability
Subject to Doubao's usage limits and terms of service
Download location may vary by system/browser configuration
Generation time varies based on server load
No guaranteed consistency in image quality or style

Integration Points

Can be combined with report generation skills (e.g., xiaoxiyouxi-5min-report)
Suitable for avatar creation, concept visualization, or illustration needs
Can feed into social media content pipelines
Useful for rapid prototyping of visual concepts

安全使用建议

This skill is inconsistent and raises red flags. Before installing, ask the publisher for: (1) a list of required automation tools (Puppeteer/Selenium, browser binary) and an install spec; (2) how Feishu (or other) delivery is authenticated and where tokens are stored; (3) confirmation they are not instructing the agent to bypass Doubao's terms or bot-detection defenses. Consider legal/ToS risk: automating to 'avoid bot detection' can violate the target site's terms and lead to account suspension or legal issues. If you still want to proceed, only run it in a controlled environment, provide minimal, well-scoped credentials (not long-lived account owner secrets), and prefer using official APIs or partnering with the service instead of automated UI scraping.

功能分析

Type: OpenClaw Skill Name: doubao-ai-image Version: 1.0.0 The skill bundle describes a legitimate workflow for automating AI image generation via the Doubao (doubao.com) web interface using browser automation. The instructions in SKILL.md focus on navigating the UI, capturing screenshots of generated images, and saving them to a local workspace directory. No indicators of malicious intent, data exfiltration, or harmful prompt injection were found; the activities are consistent with the stated purpose of providing a free alternative to image generation APIs.

能力评估

⚠ Purpose & Capability

The stated purpose (browser-automated image generation on Doubao) matches the instructions, but the skill does not declare or document required automation tooling (e.g., Puppeteer/Selenium, headless browser) or credentials needed for delivery endpoints. Asking the agent to 'mimic human interaction to avoid bot detection' is disproportionate and suggests actions beyond straightforward automation.

⚠ Instruction Scope

SKILL.md instructs browser navigation, DOM element targeting, screenshot capture, local file writes to /workspace/ai_images/doubao/, and delivering images to external chat (Feishu) — but it does not declare how the agent will obtain Feishu credentials or the browser automation runtime. The explicit guidance to avoid bot detection is scope-creep and ethically questionable; it may imply bypassing site protections or terms of service.

ℹ Install Mechanism

There is no install spec (instruction-only), which reduces file-written-to-disk risk. However, the technical requirements (browser automation, element targeting, user-agent control) imply additional tooling that the skill does not declare or install, creating an operational gap the deployer must fill.

⚠ Credentials

The skill requests no environment variables or credentials in metadata but refers to delivering images to Feishu chat and using a user agent/cookies for automation. That delivery/integration normally requires API tokens or web session credentials — their absence in the declared requirements is an inconsistency and could lead to ad-hoc credential handling or insecure storage.

✓ Persistence & Privilege

The skill is not marked always:true and does not request persistent system-wide privileges. It writes to a local workspace path per its instructions, which is expected for an image-generation task. The default ability for the agent to invoke the skill autonomously is enabled (disable-model-invocation: false), which is normal but worth noting.

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

元数据

Slug doubao-ai-image

版本 1.0.0

许可证 MIT-0

累计安装 0

当前安装数 0

历史版本数 1

常见问题