| name | shopify-blog-publisher |
| description | End-to-end Shopify blog article publisher. Handles keyword research (DataForSEO), blog writing, featured image generation (Kie AI), and publishing to Shopify — all in one flow. Use this skill whenever the user wants to publish a blog post to their Shopify store, write a blog article for Shopify, do keyword research for a blog, or says things like "publish a blog", "new blog post", "write an article for the shop", "blog for Shopify", "create a blog post", or any variation of wanting to research, write, and publish blog content to their Shopify webshop. |
Shopify Blog Publisher
Publishes a complete, SEO-optimized blog article to a Shopify store in one flow: keyword research, writing, image generation, and publishing.
Prerequisites
Before using this skill, you need:
- Shopify store with a Custom App that has
write_content scope (see README.md)
- DataForSEO account for keyword research (https://dataforseo.com)
- Kie AI MCP installed in your Claude workspace for image generation
Set your credentials as environment variables — see the .env.example file included in this skill.
The Flow
1. Keyword Research (DataForSEO)
→ Find the right topic angle and keywords
2. Blog Writing
→ Write the article optimized for the researched keywords
3. Image Generation (Kie AI)
→ Create a featured image for the post
4. Publish to Shopify
→ Push the finished article live on the store
Each phase builds on the previous one. Don't skip phases unless the user explicitly says to.
Phase 1: Keyword Research (DataForSEO)
Use the scripts/dataforseo.py script for all DataForSEO API calls. The script handles auth and response parsing.
Step 1: Get the seed keyword from the user
Ask the user for a topic or seed keyword. If they've already provided one, use it.
Step 2: Run keyword suggestions
python3 <skill-path>/scripts/dataforseo.py suggestions "<seed_keyword>" --location <location_code> --language <language_code> --limit 20
This returns keyword suggestions with search volume, CPC, and competition.
Default: location 2840 (United States), language en (English). Override per request — e.g., Netherlands = 2528 / nl, UK = 2826 / en, Germany = 2276 / de.
Step 3: Run search volume for promising keywords
Pick the 5-10 most relevant keywords from suggestions and get precise search volume:
python3 <skill-path>/scripts/dataforseo.py volume "<keyword1>,<keyword2>,<keyword3>" --location <location_code> --language <language_code>
Step 4: Run related keywords for additional angles
python3 <skill-path>/scripts/dataforseo.py related "<seed_keyword>" --location <location_code> --language <language_code> --limit 10
Step 5: Present findings to the user
Show a clean summary table:
| Keyword | Search Volume | Competition | CPC |
|---|
Recommend a primary keyword and 3-5 secondary keywords. Explain why — search volume, competition level, and relevance to their store. Let the user confirm or adjust before moving to writing.
Phase 2: Blog Writing
Write the blog article optimized for the chosen keywords. The user may have specific blog preferences configured — check with them if this is the first run.
Default blog structure (if no preferences set yet):
- Title — Include primary keyword, keep under 70 characters
- Meta description — 150-160 characters, includes primary keyword
- Introduction — Hook the reader, introduce the topic, mention primary keyword naturally
- H2 sections (3-5) — Each addresses a subtopic, weave in secondary keywords
- Conclusion — Summarize key points, include a CTA
- Tags — Relevant tags based on keywords and topic
SEO guidelines:
- Primary keyword in title, first paragraph, and at least 2 H2 headings
- Secondary keywords distributed naturally throughout
- Use internal linking placeholders:
[INTERNAL LINK: related product/collection]
- Keep paragraphs short (2-3 sentences)
- Include alt text suggestions for images
- Aim for 800-1500 words depending on topic complexity
Output format:
Present the complete article to the user in a clean format. Include:
- Suggested title
- Meta description
- Full article body (HTML-ready)
- Suggested tags
Wait for the user to approve or request changes before proceeding.
Phase 3: Image Generation (Kie AI)
Generate a featured image using the mcp__kie-ai__generate_nano_banana tool.
Step 1: Craft the image prompt
Based on the blog topic and content, create a descriptive image prompt. Consider:
- The blog's main theme and mood
- Brand style (if defined — check with user)
- What would look good as a blog header/featured image
Step 2: Generate the image
Use the MCP tool:
mcp__kie-ai__generate_nano_banana(prompt="<your detailed image prompt>")
This returns a task_id. Poll for completion:
mcp__kie-ai__get_task_status(task_id="<task_id>")
Poll every 5-10 seconds until the status is complete. The result will contain an image URL.
Step 3: Present to user
Show the generated image and ask if they're happy with it. Offer to regenerate with a different prompt if needed.
Phase 4: Publish to Shopify
Use the scripts/shopify.py script for Shopify API calls. It handles OAuth token exchange and article creation.
Step 1: Get a fresh access token
python3 <skill-path>/scripts/shopify.py auth
Tokens expire every 24 hours, so always get a fresh one before publishing.
Step 2: Create the article
python3 <skill-path>/scripts/shopify.py publish \
--token "<access_token>" \
--title "<article title>" \
--body "<html body>" \
--tags "<tag1,tag2,tag3>" \
--image-url "<featured image url>" \
--published true
Set --published false if the user wants to save as draft first.
Step 3: Confirm publication
Show the user:
- Article URL on the store
- Title and status (published/draft)
- Offer to make it a draft if they published by accident
Configuration
All configuration is done via environment variables. See .env.example for the full list.
Required Environment Variables
| Variable | Description |
|---|
SHOPIFY_STORE | Your store domain (e.g., your-store.myshopify.com) |
SHOPIFY_BLOG_ID | The blog ID to publish to (find in Shopify admin URL) |
SHOPIFY_CLIENT_ID | Custom App client ID |
SHOPIFY_CLIENT_SECRET | Custom App client secret |
DATAFORSEO_LOGIN | Your DataForSEO login email |
DATAFORSEO_PASSWORD | Your DataForSEO API password |
Optional Environment Variables
| Variable | Default | Description |
|---|
SHOPIFY_API_VERSION | 2025-01 | Shopify Admin API version |
DATAFORSEO_AUTH_TOKEN | — | Pre-encoded Base64 token (alternative to login+password) |
Default Market
- Location: United States (code 2840)
- Language: English (code en)
The user can override these per-request (e.g., "target the Netherlands market" → location 2528, language nl).
Common Location Codes
| Country | Code | Language |
|---|
| United States | 2840 | en |
| United Kingdom | 2826 | en |
| Netherlands | 2528 | nl |
| Germany | 2276 | de |
| France | 2250 | fr |
| Canada | 2124 | en / fr |
| Australia | 2036 | en |
Error Handling
- DataForSEO fails: Check if the API returns status_code != 20000. Show the error to the user.
- Shopify token expired: Re-run
shopify.py auth to get a fresh token.
- Shopify publish fails: Check the HTTP status and error message. Common issues: missing required fields, HTML validation errors.
- Kie AI generation fails: Retry with a simplified prompt. If persistent, let the user know and offer to publish without an image.
- Missing credentials: The scripts will tell you exactly which environment variables are missing.
- Network errors: Retry once, then inform the user.
