AEM AI资源列表

Awesome AEM AI

A curated list of AI resources, tools, MCP servers, agents, and skills for Adobe Experience Manager (AEM) and Edge Delivery Services (EDS).

The AEM AI landscape moves fast — agents, MCP servers, and skills ship monthly across Experience League, aem.live, developer blogs, and press releases. This list is the single map, refreshed as things change.

Contents

• Start Here
• What's New (October 2026)
• Official Resources
• Architecture
• MCP Servers
• AI Agents
• Claude Code Skills
• IDE Extensions & Plugins
• AI Features by AEM Product
• Adobe LLM Optimizer
• Edge Delivery Services AI
• Development Tools
• Real-World Architecture Audit
• Best Practices & AI Workflow
• Tutorials & Learning
• Community Projects
• Videos & Presentations
• Related Awesome Lists

Start Here

Things this list has that you won't find in the official docs:

What's New (October 2026)

The current feature release is 2026.9.0 (September 24, 2026); 2026.10.0 is planned for October 29, 2026.

• AEM's agentic capabilities are now GA inside Adobe CX Enterprise Coworker (see AI Agents below). Since early September they run as part of Coworker - an AI teammate for marketing and CX teams - reached through the Ask Coworker entry point rather than standing alone in AEM.
• The agent documentation restructured into Agentic Capabilities, grouped as Brand Experience (Experience Production, Development, Onboarding), Content Advisor (Discovery, Content Optimization), and Brand Governance. The older per-agent pages are still published alongside them.
• The unified AEM MCP Server absorbed Cloud Manager operations - programs, environments, repositories, and pipelines - in 2026.9.0. The domain-specific MCP servers are now deprecated; switch to the single aem endpoint.
• IDE agent skills gained AEM query optimization: as of October 1 the Code Assessment Agent Skill checks whether a JCR/Oak query is actually served by an index, and flags index definitions that look correct but are subtly misconfigured. The Cloud Migration Agent Skill added more AEM 6.5-to-Cloud-Service code migration patterns.
• Permission management went agentic in 2026.8.0 (GA) - it audits effective ACLs and explains access decisions in real time.
• Content Hub gained AI Search and Smart Collections, and Dynamic Media with OpenAPI gained AI-generated video captions - all three in Limited Availability, not yet GA. In beta: the Content Supply Chain Agent (CSC Blueprint plus CSC Integration), CX Coworker developer troubleshooting for CDN/replication/workflows, Dynamic Media Auto Translate, and the Content Advisor browser extension.
• A new AEM Content AI documentation set (linked under Official Resources below) positions governed AEM content as the ground truth that AI experiences and agents are grounded in.

Official Resources

Adobe Documentation

• AI in AEM Overview - All AI capabilities in AEM.
• Using MCP with AEM as a Cloud Service - Official MCP integration guide.
• Local Development with AI Tools - Official guide for local AI-enhanced development.
• AI-Assisted Development - Learn tutorials for agent skills, AGENTS.md, and MCP setup.
• Agents in AEM (Learn) - Hands-on tutorial for AEM agents.
• AEM MCP Modules for Workfront Fusion - Call AEM MCP tools from Fusion automation scenarios.

Adobe AI Products

• Adobe LLM Optimizer - Generative Engine Optimization (GEO) for AI search visibility.
• Adobe Brand Concierge - AI-powered conversational companion for websites, with AEM as a content source.
• AEM Content AI - Semantic search, grounded generation, and APIs that make governed AEM content the ground truth for AI experiences.
• Adobe GenStudio - AI-powered content supply chain for performance marketing.
• Content Hub - AI-enhanced asset management.
• AI Assistant in AEM - Integrated AI assistant for product knowledge and support.

Announcements

• Adobe Brand Visibility Solution (April 2026) - Summit 2026 introduced AEM as the "brand context layer," LLM Apps for building experiences inside LLM interfaces, and agentic authoring.
• The Agentic Evolution of Adobe Experience Manager (February 2026) - MCP/A2A APIs, instructional authoring, in-product agent, and AEM Playground.
• Adobe AI Agents GA (September 2025) - General availability of AI agents for customer experience orchestration.

Architecture

Adobe Experience Platform Agent Orchestrator

The Agent Orchestrator is the agentic layer powering AEM AI agents. It coordinates specialized agents across workflows and applications.

┌─────────────────────────────────────────────────────────────────────┐
│                    AI Assistant (Conversational UI)                  │
├─────────────────────────────────────────────────────────────────────┤
│                         Reasoning Engine                             │
│   • Interprets natural language prompts                              │
│   • Creates step-by-step execution plans                             │
│   • Adjusts dynamically, retries alternative approaches              │
├─────────────────────────────────────────────────────────────────────┤
│                    Adobe Experience Platform Agents                  │
│  ┌──────────────┐ ┌──────────────┐ ┌──────────────┐ ┌─────────────┐ │
│  │  Discovery   │ │   Content    │ │  Experience  │ │ Development │ │
│  │    Agent     │ │ Optimization │ │  Production  │ │   Support   │ │
│  └──────────────┘ └──────────────┘ └──────────────┘ └─────────────┘ │
├─────────────────────────────────────────────────────────────────────┤
│                         Knowledge Base                               │
│   • Adobe product documentation                                      │
│   • Customer metadata & business objects                             │
│   • Analytics data                                                   │
└─────────────────────────────────────────────────────────────────────┘

Documentation: Platform Learn Tutorial

Key Components

Built-in Functional Agents

Experience Modernization Agent Architecture

The Experience Modernization Agent (documented in the AI Agents section below) automates migration to Edge Delivery Services:

┌─────────────────────────────────────────────────────────────────────┐
│                   Experience Modernization Console                   │
│                      (Web Interface)                                 │
├─────────────────────────────────────────────────────────────────────┤
│                        AI Coding Agent                               │
│  ┌──────────────────────────────────────────────────────────────┐   │
│  │                    Migration Skills                           │   │
│  │  • Page decomposition & section identification                │   │
│  │  • Block mapping against library                              │   │
│  │  • Design system extraction (colors, fonts, CSS)              │   │
│  │  • Bulk content import                                        │   │
│  └──────────────────────────────────────────────────────────────┘   │
├─────────────────────────────────────────────────────────────────────┤
│                      Integration Layer                               │
│  ┌─────────────┐  ┌─────────────┐  ┌─────────────┐  ┌────────────┐  │
│  │   GitHub    │  │    Figma    │  │  Live AEM   │  │   Source   │  │
│  │ Integration │  │ MCP Server  │  │   Preview   │  │  Website   │  │
│  └─────────────┘  └─────────────┘  └─────────────┘  └────────────┘  │
├─────────────────────────────────────────────────────────────────────┤
│                    Edge Delivery Services                            │
│              (Production-ready output)                               │
└─────────────────────────────────────────────────────────────────────┘

Workflow:

1. Analyze source pages → Identify visual sections
2. Map sections → Match against block library
3. Extract design system → Generate CSS (colors, fonts, styles)
4. Create PRs → GitHub integration for review
5. Deploy → Direct to Edge Delivery Services

Security & Governance

Documentation: AI in Experience Cloud

Core AI Services & Trust

Foundational AI technologies and ethics powering the AEM ecosystem.

Architecture Resources

MCP Servers

Model Context Protocol (MCP) servers that integrate with AEM and EDS.

AEM MCP Server (Unified, Recommended)

Adobe recommends a single endpoint over configuring the domain-specific servers individually. As of 2026.9.0 it also covers Cloud Manager operations, which previously required the separate server:

This same endpoint backs the AEM Claude Connector (install from Claude's connector directory on claude.ai) and the AEM ChatGPT Plugin (one-click install from OpenAI's plugin marketplace).

Connecting a client: Adobe documents setup per tool in the Setup Overview - Claude and ChatGPT as above, plus Cursor, JetBrains with GitHub Copilot, and Microsoft Copilot Studio.

Official Adobe MCP Servers (Domain-Specific, Deprecated)

Deprecated as of 2026.9.0. They still work, but Adobe recommends the unified AEM MCP Server above instead - Cloud Manager operations in particular moved there. Cloud Migration is the exception: it is not part of the unified endpoint yet, so it stays in use. Hosted at https://mcp.adobeaemcloud.com/adobe/mcp/:

Key Features:

• Natural language interaction with AEM content
• OAuth authentication via Adobe ID
• Respects user's AEM permissions
• Supports Claude, ChatGPT, Cursor, VS Code, JetBrains, GitHub Copilot, Claude Code, Cline, Windsurf, and Microsoft Copilot Studio

See the official MCP documentation linked under Official Resources above.

Local Runtime MCP Servers

Model Context Protocol (MCP) servers that connect AI tools to local running environments (AEM SDK and Dispatcher).

See the Local Development with AI Tools guide linked under Official Resources above.

Community MCP Servers

Edge Delivery Services MCP Servers

Also recommended by the aem.live AI coding agents guide: Context7 (API docs indexing including AEM), DA MCP Server (Document Authoring content workflows), and Browser MCP (remote browser control and screenshots).

Official Adobe MCP Tutorials

Adobe's own end-to-end walkthroughs for the AEM MCP Server on Experience League.

AI Agents

Official Adobe Agents

Adobe's AEM agents are organized into three top-level agents. Overview

Try before you buy: AEM Playground - isolated sandbox for agentic workflows (auto-deletes after 30 days). Production access via the Agentic SKU - contact your Adobe CSM/TAM.

Where the Agents Run: CX Enterprise Coworker

Since early September 2026, AEM's agentic capabilities are generally available inside Adobe CX Enterprise Coworker, an AI teammate for marketing and CX teams. Authors reach them through the Ask Coworker entry point, and the capabilities collaborate with Coworker Chat rather than running as standalone agents inside AEM.

Rollout caveat from the release notes: once enabled, only the Coworker immersive experience is available - the in-product right-rail panel inside AEM comes later.

Adobe now also documents the AEM side under Agentic Capabilities, grouped differently from the per-agent pages below. Both doc trees are live:

Brand Experience Agent

Umbrella agent for operational automation. Overview

Experience Production Agent

Automates high-effort, high-volume content tasks. Overview

Content Update Skill Technical Details:

Experience Modernization Agent

AI-powered website migration to Edge Delivery Services. Overview

Console: Available at aemcoder.adobe.io - no local setup required

Experience Development Agent

AI-assisted troubleshooting and build automation for AEM Cloud Service. Overview

Access: Requires Cloud Manager Developer or Program Manager role

Content Advisor Agent

Discover, refine, and adapt assets through natural language. Overview

Governance Agent

Brand integrity and compliance enforcement. Overview

Integration: Works with ChatGPT, Claude, and other AI systems via A2A and MCP protocols, plus the official /experience-governance MCP server

Availability: AEM as a Cloud Service and Edge Delivery Services (GA via Agentic SKU); progressive rollout to AEM 6.5 LTS on Managed Services is planned

Other Agentic Capabilities

Capabilities that do not sit under the three top-level agents:

Claude Code Skills

Official Adobe Skills (adobe/skills)

The adobe/skills repository is production-ready (main branch) and ships as a plugin marketplace. The AEM plugins are aem-cloud-service, aem-6-5-lts, aem-edge-delivery-services, aem-eds-content-ops, and aem-project-management; the repo also carries App Builder, Commerce, Analytics/CJA, Workfront, and Creative Cloud plugins.

# Add the marketplace, then install the plugin you want (recommended)
/plugin marketplace add adobe/skills
/plugin install aem-cloud-service@adobe-skills

# Install via npx (Vercel Skills)
npx skills add adobe/skills --all

# Install via gh-upskill
gh upskill adobe/skills --all

AEM Cloud Service skills (aem-cloud-service):

AEM 6.5 LTS skills (aem-6-5-lts): dispatcher, aem-workflow, aem-replication, ensure-agents-md, and tuning-oak-query-indexes.

EDS content operations (aem-eds-content-ops) adds 13 content and SEO skills: accessibility-fix, bulk-metadata, content-audit, content-diff, cwv-optimizer, geo-rewrite, heading-optimizer, image-seo, internal-linking, link-rot-scanner, product-page-seo, sitemap-audit, structured-data.

Official Adobe EDS Skills

Install official Adobe skills from adobe/helix-website:

# Install via gh-upskill
gh extension install trieloff/gh-upskill
gh upskill adobe/helix-website --all

15 skills available: content-driven-development, page-import, building-blocks, content-modeling, testing-blocks, code-review, docs-search, block-inventory, block-collection-and-party, page-decomposition, identify-page-structure, scrape-webpage, generate-import-html, preview-import, authoring-analysis

AEM Core Skills (This Repository)

Pre-built AEM backend skills available in .claude/skills/:

Installation

# Copy this repo's skills to your AEM project
cp -r .claude/skills/ /path/to/your/aem-project/.claude/skills/

Skills Tools

Documentation

• Claude Code Skills Docs - Official skill creation guide.
• Agent Skills Repository - Example skills and the skill specification.

IDE Extensions & Plugins

VS Code Extensions

IntelliJ Plugins

AI Features by AEM Product

Adobe GenStudio

See the GenStudio product page linked under Official Resources above.

AEM Sites

AEM Assets

AEM Forms

AEM Guides

AEM Screens

Note: AEM Screens inherits AI capabilities from AEM Sites and Assets (Smart Tags, Smart Crop, etc.) for content that is published to digital signage displays.

Migration & Modernization

Adobe LLM Optimizer

Generative Engine Optimization (GEO) for AI search visibility in ChatGPT, Perplexity, Copilot, Gemini, and other LLM-driven assistants.

Overview

Adobe LLM Optimizer is a generative AI-first application designed to help brands enhance their visibility, accuracy, and influence in AI-driven search environments. It provides insights into brand presence in AI-generated answers, offers prescriptive content recommendations, and automates optimization fixes.

Key Features

LLM Optimizer is part of Adobe's Brand Visibility solution (Summit 2026, linked under Announcements above) alongside Brand Concierge and LLM Apps.

Tutorials & Guides

Integration

• Native integration with AEM as a Cloud Service
• Optimize EDS sites for LLM visibility
• Configure CDN for AI agent traffic targeting

Use Cases:

• Improve brand citations in AI-generated answers
• Optimize content structure for LLM readability
• Track performance across ChatGPT, Google AI Overviews, Copilot, Gemini, Perplexity
• Automated optimization fixes via CDN edge deployments

Getting Started

1. Complete domain onboarding
2. Configure categories, topics, and prompts
3. Set up CDN log forwarding
4. Monitor brand presence dashboard
5. Apply optimization recommendations

The official documentation home is linked under Official Resources above.

Edge Delivery Services AI

Official Documentation (aem.live)

Adobe Helix Claude Skills

Official skills from the adobe/helix-website repository (linked under Claude Code Skills above):

Installing Adobe Skills

# Install gh-upskill extension
gh extension install trieloff/gh-upskill

# Add Adobe EDS skills to your project
gh upskill adobe/helix-website

# Or install standalone
curl -fsSL https://raw.githubusercontent.com/trieloff/gh-upskill/main/install.sh | bash
upskill adobe/helix-website --all

Development Tools

AI Grounding & Configuration

Essential files for grounding AI agents (Claude, Cursor, Copilot) in AEM Cloud Service project context.

AEM Code Assessment IDE Agent

For AI-assisted IDEs (Cursor, Claude Code, Visual Studio, IntelliJ): detects and auto-fixes deprecated API usage, outdated Sling Model dependency injection, outdated Maven dependencies, missing HTTP call timeouts, and unbounded queries/schedulers/listeners (2026.7.0). As of 2026.9.0 it also performs AEM query optimization, flagging JCR/Oak index misconfiguration - GA on October 1, 2026. See the current release notes linked under What's New above.

npm Packages

Official Adobe Packages

Development Tools

CLI Commands

# AEM CLI - local development server with AI coding agents support
aem up

# AEM Project Archetype
mvn archetype:generate -DarchetypeGroupId=com.adobe.aem -DarchetypeArtifactId=aem-project-archetype

Real-World Architecture Audit

To demonstrate the difference between "GenAI chat" and "Architectural Intelligence," we performed a live audit of the AEM WKND Reference Site Dispatcher configurations using the principles defined in this guide.

The Objective

Identify security vulnerabilities in the Dispatcher filter rules that a standard chatbot would miss, but a grounded AEM agent can detect.

Step 1: Automated Analysis

We ran the aemanalyser-maven-plugin (integrated via the dispatcher skill) against the wknd-test project.

# Executing the audit
mvn com.adobe.aem:aemanalyser-maven-plugin:analyse -pl dispatcher -am

Step 2: Findings (Critical Vulnerability Detected)

The audit identified a high-risk misconfiguration in dispatcher/src/conf.dispatcher.d/filters/filters.any:

Vulnerable Code:

# Rule /0201 in filters.any
/0201 { /type "allow" /url "/home/users/*.infinity.json" }

Architectural Impact:

Step 3: Proactive Fix

A "really good" AI doesn't just report this; it provides the surgical fix:

# Recommended Fix: Restrict access to specific, non-sensitive pagedata only
/0201 { /type "deny" /url "/home/users/*" }
/0202 { /type "allow" /url "/home/users/*.token.json" } # If specific token exchange is needed

Best Practices & AI Workflow

Guidelines for successfully integrating AI into your AEM development lifecycle.

Core Principles

Standard AI-Enhanced Workflow

Tutorials & Learning

Official Adobe Learning

Hands-on Labs

Step-by-step tutorials for practical AI implementation in AEM.

Quick Start:

git clone https://github.com/narendragandhi/awesome-aem-ai.git
cd awesome-aem-ai/tutorials

Certifications

• Adobe Certification Program - AEM Sites, Assets, and Developer certifications.

Community Projects

Open Source

Community Tools

Videos & Presentations

Adobe Developers Live

AI & MCP Tutorials

YouTube Channels

• Adobe Experience League - Official Adobe tutorials and release updates.
• AEM Geeks - Community AEM tutorials.
• aem-live - Official Adobe AEM channel.

Webinars

• Monthly AEM Community Webinars - Recurring community sessions on Experience League.

Related Awesome Lists

• Awesome AEM - Curated list of AEM/CQ5 resources.
• Awesome MCP Servers - MCP server collection.
• Awesome LangChain - LangChain resources.
• Awesome Claude Code - Claude Code skills, hooks, and tools.

Contributing

Contributions are welcome! Please read the contribution guidelines first.

How to Contribute

1. Fork this repository
2. Create a new branch (git checkout -b feature/add-resource)
3. Add your resource to the appropriate section
4. Commit your changes (git commit -am 'Add awesome resource')
5. Push to the branch (git push origin feature/add-resource)
6. Create a Pull Request

Contribution Guidelines

• Ensure the resource is relevant to AEM and AI
• Provide accurate descriptions
• Include working links
• Follow the existing format

_{Built with AI for the AEM Community · Licensed under CC0 1.0}