菜单
Generate images using ChatGPT (DALL-E) via Chrome browser automation. Use when the user wants to create, generate, or produce an image using ChatGPT. Requires the Claude-in-Chrome browser extension to be active. Invoked as /chatgpt-image with an image description.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/rawe/agentic-coding-guide --skill chatgpt-image命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
正在显示 SKILL.md
基于 SOC 职业分类
| name | chatgpt-image |
| description | Generate images using ChatGPT (DALL-E) via Chrome browser automation. Use when the user wants to create, generate, or produce an image using ChatGPT. Requires the Claude-in-Chrome browser extension to be active. Invoked as /chatgpt-image with an image description. |
| argument-hint | ["image description"] |
| user-invocable | true |
Generate an image via ChatGPT/DALL-E using Chrome browser automation, download it, and save it to the project's generated-images/ directory with a descriptive filename.
Given an image description via $ARGUMENTS:
Get browser context: Call tabs_context_mcp to discover available tabs.
Open ChatGPT: Check if a chatgpt.com tab exists. If yes, reuse it. If not, create a new tab and navigate to https://chatgpt.com.
Enter prompt: Click the input field ("Ask anything"), type:
Generate an image: <$ARGUMENTS>
Submit: Click the submit button (circular arrow icon, bottom-right of input field).
Wait for generation: Take screenshots every 5–8 seconds to monitor. The image is ready when the "Creating image" placeholder is replaced by the actual image with "Image created" label. Timeout after 60 seconds if not complete.
Download the image: Use read_page with filter: "interactive" to find the button labeled "Download this image" in the accessibility tree. Click it via its ref ID. For detailed interaction specifics, see references/browser-workflow.md.
Wait for download: Wait 3 seconds for the file to appear in ~/Downloads/.
Move and rename: Derive a filename slug from $ARGUMENTS (lowercase, hyphens, no special chars, max 60 chars, .png extension). Run:
.claude/skills/chatgpt-image/scripts/move-chatgpt-image.sh generated-images/<slug>.png
Confirm: Report the saved path to the user.