refactor: restructure mytoolkit with new subcommands and md-to-pdf improvements

- Add new subcommands: convert, preflight, server, templates, webpage
- Migrate config from bin/config.json to ~/.config/mytoolkit
- Fix expand_bookmarks to modify writer objects instead of reader
- Improve md_to_pdf with CJK bookmark support
- Update README and project metadata

bin/commands/image.py

@@ -1,10 +1,14 @@
 """Image utilities."""
 
+import base64
+import os
 import subprocess
+import urllib.request
 from pathlib import Path
 
 import click
 
+from bin.config import config
 from bin.utils import handle_errors, run_command
 
 
@@ -181,3 +185,52 @@ def compress_images(files, target_mb, max_width, quality, suffix, dry_run):
                     size_mb = output.stat().st_size / (1024 * 1024)
 
             click.echo(f"Compressed: {img_file} -> {output} ({size_mb:.2f} MB)")
+
+
+@image.command("generate")
+@click.argument("prompt", required=False)
+@click.option("-f", "--file", type=click.Path(exists=True), help="从文件读取 prompt")
+@click.option("-s", "--size", default="2K", type=click.Choice(["1K", "2K"]), help="图片尺寸")
+@click.option("-o", "--output", help="保存路径（默认只输出 URL）")
+@click.option("--no-watermark", is_flag=True, help="不添加豆包水印")
+@click.option("--b64", is_flag=True, help="使用 b64_json 格式获取图片数据")
+@handle_errors
+def generate_image_cmd(prompt, file, size, output, no_watermark, b64):
+    """使用豆包/Seedream 模型生成图片。"""
+    from openai import OpenAI
+
+    if file:
+        prompt = Path(file).read_text(encoding="utf-8").strip()
+    elif not prompt:
+        raise click.UsageError("必须提供 prompt 或使用 -f/--file 从文件读取")
+
+    api_key = config.get("apikey_ark")
+    if not api_key:
+        click.echo("Error: apikey_ark not set. Run: mytoolkit env set apikey_ark <value>", err=True)
+        raise click.Abort()
+
+    client = OpenAI(
+        base_url="https://ark.cn-beijing.volces.com/api/v3",
+        api_key=api_key,
+    )
+
+    resp = client.images.generate(
+        model="doubao-seedream-5-0-260128",
+        prompt=prompt,
+        size=size,
+        response_format="b64_json" if b64 else "url",
+        extra_body={"watermark": not no_watermark},
+    )
+
+    if b64:
+        image_bytes = base64.b64decode(resp.data[0].b64_json)
+        save_path = output or "generated_image.png"
+        Path(save_path).write_bytes(image_bytes)
+        click.echo(os.path.abspath(save_path))
+    else:
+        url = resp.data[0].url
+        if output:
+            urllib.request.urlretrieve(url, output)
+            click.echo(os.path.abspath(output))
+        else:
+            click.echo(url)