Aeko setup store
Self-setup flow for AEKO through MCP. Adds a domain, connects a Cafe24/Shopify store or injects products for a custom/manual store, lists products, updates target markets, checks quota, and generates/accepts starter tracked prompts. Starter-compatible for core setup; Pro+ features such as Context Reviews, marketing ads, content recommendations, and aeko.shop publishing stay separate.From its SKILL.md
npx -y skills add AEKO-Intelligence/aeko-plugin --skill aeko-setup-storeAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
SKILL.md
5.6 KB, as published. Nobody here has run it
AEKO Setup Store
Use this when a user wants to onboard AEKO end-to-end through the agent, including custom/AWS-built stores that do not have a Cafe24 or Shopify connector.
Language: mirror the user's chat language for user-facing questions and summaries. Keep tool names, IDs, country codes, product keys, and JSON field names in English/ASCII.
Tier Contract
Starter can:
- add domains within the Starter domain cap
- add/track prompts within the Starter prompt cap
- use one selected market
- use OpenAI as the allowed AI platform
- connect/sync a store or inject products for a manual/custom store
- list products and perform store-write workflows that the backend allows for active subscriptions
Pro+ is required for:
- Context Reviews / review integrations (connect Crema/Judge.me/Cafe24 in the AEKO dashboard — review-source connection is not available through the agent; you can still inject reviews for a custom store via
/aeko-inject-reviews) - Context library grounding
- content recommendations/content-plan generation
- OpenAI Ads marketing workspace
- aeko.shop publishing
If the user asks why a core setup action is blocked for Starter, treat that as a bug or quota issue unless the backend error explicitly says the feature is Pro+.
Step 1 - Resolve or Add Domain
Call aeko_list_domains.
If the user passed $1 and it looks like a URL, add it with:
aeko_add_domain(base_url=<url>, display_name=<optional>, scope=<optional>, ko_name=<optional>)
If no domains exist, ask for the store/domain URL and then call aeko_add_domain.
If multiple domains exist, show the IDs and ask which one to configure. If one domain exists, auto-select it unless the user supplied a different URL.
After selecting/adding, call aeko_get_domain_info(domain_id) and summarize the chosen domain.
Step 2 - Select Store Path
Ask which path applies:
- Cafe24
- Shopify
- Custom/manual store
For Cafe24/Shopify:
- Ask for
store_identifierand credentials/tokens only if the user already has them available. - Call
aeko_connect_store(domain_id, platform, store_identifier, access_token, ...). - Then call
aeko_sync_store(integration_id)from the returned store. - If the user does not have credentials, say browser OAuth/dashboard connection is required for now and stop store sync cleanly.
For custom/manual:
- Ask the user for product data or a CSV/table they can paste.
- Never fabricate products.
- Require each product to have:
external_product_idtitleproduct_urlpublic_url
- Recommended optional fields:
descriptionpricecurrencyimage_urlstatus(sellingby default)
- Call
aeko_inject_products(domain_id, products); the backend creates the manual store internally.
Step 3 - Verify Products
Call:
aeko_list_store_products(domain_id=<domain_id>, include_citability=false, limit=50)
Show product count and a compact table:
| external_product_id | title | status | public URL |
|---|---|---|---|
If zero products return after a sync/inject, stop and explain the likely reason from the tool output.
Step 4 - Markets
Ask which target markets to monitor.
Starter: choose one market. Pro/Enterprise: multiple markets are allowed within backend limits.
Call:
aeko_update_markets(["US"])
If the backend rejects the market count, surface that message and ask for a smaller set.
Step 5 - Quota
Call aeko_get_quota.
Summarize:
- tracked prompts used
- remaining tracked prompt slots
- domain/market limits if returned
Do not attempt starter prompt acceptance if quota is already full.
Step 6 - Starter Prompts
Call:
aeko_generate_starter_prompts(domain_id)
Show the generated prompts with reason/grade if present. Ask which prompts to accept.
After explicit selection, call:
aeko_accept_starter_prompts(domain_id, selections)
Selections must preserve the generated prompt fields the backend expects:
raw_promptprompt_kindtarget_market- optional
prompt_en - optional score/grade fields
- optional
ai_platforms/countries
For Starter, use one allowed market and OpenAI unless the backend/domain data says otherwise.
Step 7 - Summary
Print:
Setup complete
Domain: <domain_id> · <base_url>
Store: <manual/cafe24/shopify> · <integration_id if known>
Products available: <count>
Markets: <markets>
Starter prompts accepted: <count>
Next:
/aeko-visibility-report <domain_id>
/aeko-find-prompts-to-track <domain_id>
If the user is Pro+, optionally mention review/Context/ads follow-ups. For Starter, do not pitch Pro+ unless they attempted a Pro+ feature and received a backend gate.
Never
- Never create fake products or reviews.
- Never ask Starter users to use Context Reviews, Context library, ads, content recommendations, or aeko.shop publishing as a setup prerequisite.
- Never publish content.
- Never store credentials outside the MCP call.