From aeko-plugin
Top-level router for AEKO optimization work. Lists pending items for a domain grouped into three categories (Technical / PDP update / Content generation) and prints ready-to-copy commands that route the user to `/aeko-fix-technical`, `/aeko-update-pdp`, or `/aeko-create-content`. Pure dispatcher — never executes items itself.
How this skill is triggered — by the user, by Claude, or both
Slash command
/aeko-plugin:aeko-action-center [domain-id] [category][domain-id] [category]This skill is limited to the following tools:
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
Router for three execution categories: **Technical fixes**, **상품 페이지 개선 (PDP update)**, and **Content generation**. You help the user pick one pending item and hand off to the correct executor skill. You do NOT generate artifacts, call write-back tools, or mark items complete.
Router for three execution categories: Technical fixes, 상품 페이지 개선 (PDP update), and Content generation. You help the user pick one pending item and hand off to the correct executor skill. You do NOT generate artifacts, call write-back tools, or mark items complete.
Contract reference: docs/contracts/action-item-contract.md.
This is the front door for non-technical ecommerce marketers. Open with what AEKO found, why it matters
for AI visibility or shopping discovery, and the safest next command. Do not show execution_class,
contract versions, raw IDs, or backend jargon unless needed for debugging. Group work as:
Technical health, Product pages, and Content AI can cite.
Language: mirror the user's chat language for user-facing steps, summaries, questions, and next actions. Keep slash commands, IDs, file paths, channel slugs, schema keys, and tool names in English/ASCII.
domain-id (optional) — UUID of the domain. If missing, ask the user (or call aeko_list_domains to offer a pick-list).category (optional) — technical | pdp | content | all (default: all).Parse $1 for a UUID. If absent:
aeko_list_domains to show the user's connected domains and offer a pick-list, ORaeko_get_domain_info with a guess if one is obvious from context.Do not proceed without a concrete domain_id.
Call the two list endpoints in parallel (fast, independent):
aeko_list_action_items(domain_id, status="pending,ready") → returns items with tab="action".aeko_list_technical_items(domain_id, status="pending,ready") → returns items with tab="technical".Each summary (AekoItemSummary) carries: id, tab, title, priority, execution_class, artifact_type, write_mode, preview, updated_at, target_url, product_id.
If either endpoint errors, report it plainly and continue with whichever returned data. Do NOT fall back to retired tools (aeko_get_suggestions_v2, aeko_get_suggestions, campaigns). Those are deleted in v0.5.0 — any reference means stale skill code.
Group the combined item list into three buckets:
| Category | Filter | Executor |
|---|---|---|
| Technical health (기술 상태) | execution_class == "technical_artifact" | /aeko-fix-technical <item_id> |
| Product pages (상품 페이지) | execution_class == "store_write_artifact" | /aeko-update-pdp <item_id> |
| Content AI can cite (인용 가능한 콘텐츠) | execution_class == "local_content_artifact" | /aeko-create-content <item_id> |
If $2 (category arg) is set to technical / pdp / content, filter to only that bucket. all or unset → show all three.
Domain: <brand_or_domain_name> (<base_url>)
Technical health: N pending (top priority: <critical|high|medium|low>)
Product pages: M pending (top priority: ...)
Content AI can cite: K pending (top priority: ...)
Before item blocks, add one plain-language sentence:
Recommended next step: <one command> — <why this is the safest/highest-impact next move>.
For each category with ≥1 pending item, group items by priority (critical → low) and print a fenced code block per item, highest-priority-first. Cap at 10 items per category; if there are more, append a line noting how many are hidden.
Technical items:
# Technical health · <priority>
# <title>
# Fix type: <plain-language label, e.g. "AI crawler access" or "site schema">
# updated <relative time>
# preview: <first 120 chars of preview>
/aeko-fix-technical <item_id>
PDP update items:
# Product page · <priority>
# <title>
# Safety: <preview only | shadow copy | live store update after confirmation>
# updated <relative time>
# target: <target_url or product_id>
# preview: <first 120 chars of preview>
/aeko-update-pdp <item_id>
Content generation items:
# Content AI can cite · multi-channel · <priority>
# <title>
# asks you to choose channels, then drafts each selected channel
# updated <relative time>
# preview: <first 120 chars of preview>
/aeko-create-content <item_id>
The multi-channel tag and the fan-out hint are required — they signal that running the executor opens an interactive channel-selection flow (auto-detected channels from cited-source analysis + optional addons: press_release / magazine / instagram / tiktok / youtube / other:<name>) and is not a one-shot generator. Do not omit either line.
After printing, ask which one the user wants to tackle. They copy the command block and run it themselves.
Refuse: tell the user to run each executor one item at a time so they can review the output between runs. Writes to the store and content artifacts should not batch.
domain_id missing AND aeko_list_domains returns zero → tell the user to add a domain in the AEKO dashboard first; stop./aeko-brand-kit <domain_id>) or running /aeko-visibility-report <domain_id>.aeko_get_action_plan — that belongs to the executor skills.aeko_complete_action_item or any write-back tool.execution_class raw to the user — always translate to the category label (Technical health / Product pages / Content AI can cite).npx claudepluginhub aeko-intelligence/aeko-plugin --plugin aeko-pluginGuides creation, editing, and verification of skills for AI coding agents using test-driven development with subagent scenarios. Use when authoring or debugging skills.