qiaomu-blog-publish

qiaomu-blog-publish: Publish content to Qiaomu Blog

Triggering

Slash command:

• /qiaomu-blog-publish path/to/file.md
• /qiaomu-blog-publish https://example.com/article
• /qiaomu-blog-publish Then paste Markdown or the body text

Natural language:

• "Publish to the blog"
• "Publish as a draft"
• "Post to Qiaomu Blog"
• "Post this article to my own blog"
• “publish to blog”
• “publish draft”

Configuration

API Base URL: https://your-domain.com

Token reading priority:

1. Environment variable QIAOMU_BLOG_API_TOKEN
2. Environment variable QMBLOG_API_TOKEN(compatible with old configuration)
3. Config file ~/.claude/skills/qiaomu-blog-publish/config.json

If there is no token, prompt the user:

1. Open https://your-domain.com/admin/settings
2. On the API Token page, generate a token
3. Save to:

{
  "apiUrl": "https://your-domain.com",
  "token": "qm_xxx"
}

Workflow

1. Determine the input source

• File path: read a local Markdown or text file
• URL: fetch the body text and convert it to Markdown
• Plain text: use directly as the body text

2. Read configuration

Prefer reading environment variables; if none, then read config.json in apiUrl and token。

3. Fetch categories

curl -s "https://your-domain.com/api/admin/categories" \
  -H "Authorization: Bearer $TOKEN"

Let the user choose a category; if they don't choose, it can be left blank and adjusted later in the backend.

4. Parse content

Title priority:

1. YAML frontmatter's title
2. The first # Heading
3. File name

Body text processing:

• Remove frontmatter
• If the title comes from the first # Heading, then remove this heading to avoid duplication
• Keep the rest of the Markdown as is

5. Upload local media

These references need to be recognized:

• ![alt](./image.png)
• ![[image.png]]
• Local paths for audio / video / attachments

Upload API:

curl -s -X POST "https://your-domain.com/api/uploads" \
  -H "Authorization: Bearer $TOKEN" \
  -F "file=@/absolute/path/to/file"

After getting the returned URL, replace the local references in the body text.

6. Transfer third-party images

If there are third-party remote images in the Markdown:

1. First download to a temporary file
2. Then upload to the blog
3. Replace with the blog's own image URL

If the download fails, keep the original image URL and indicate it in the final result.

7. Let the user confirm the publish parameters

Confirm three things:

1. Title
2. Category
3. Status:draft or published

Default to draft。

8. Publish

curl -s -X POST "https://your-domain.com/api/posts" \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "title": "The Title",
    "content": "Full processed Markdown content",
    "category": "selected-category-or-empty",
    "status": "draft"
  }'

9. Output the result

Published successfully!

Title: xxx
Status: draft
Category: xxx
Edit: https://your-domain.com/editor?slug=2026-04-16-abc123
View: https://your-domain.com/posts/2026-04-16-abc123
Files: Uploaded N files

Error handling

• No token: prompt to generate one in the backend
• 401: prompt that the token is invalid or incorrect
• Upload failed: keep the original references and indicate the failed items
• URL fetching failed: prompt the user to paste the body text directly
• Empty content: terminate and ask for additional content

Notes

• Publishes as a draft by default
• Supports uploading common files such as images, audio, video, PDF, Epub, etc.
• Suitable for use together with the ecosystem/chrome-clipper、ecosystem/obsidian-publisher in this repository