A reusable skill for generating AWS architecture diagrams in draw.io format. Works with any AI assistant — ChatGPT, Claude Projects, Kiro CLI, Claude Code, or any internal AI agent that supports custom prompts and file uploads.

Generates production-quality .drawio files using official AWS Architecture Icons with proper styling, layout, and color coding.

• Left-to-right flow — UI/Frontend on left, data sources on right
• Official AWS icons — from draw.io's built-in mxgraph.aws4 stencil library (sourced from AWS Architecture Icons)
• Verified icon catalog — only icons confirmed to render correctly
• Consistent styling — 78px icons, strokeWidth=2 edges, proper AWS color palette
• Export support — PNG/SVG/PDF via draw.io Desktop CLI with embedded XML

npx skills add vidanov/aws-architecture-diagram-skill

Install to a specific agent:

npx skills add vidanov/aws-architecture-diagram-skill --agent claude-code
npx skills add vidanov/aws-architecture-diagram-skill --agent cursor
npx skills add vidanov/aws-architecture-diagram-skill --agent kiro

Install globally (available in all projects):

npx skills add vidanov/aws-architecture-diagram-skill --global

No code or CLI required. Works with any AI that supports custom instructions and file uploads:

1. Copy the prompt from chatgpt/PROMPT.md into your assistant's system prompt / custom instructions field
2. Upload the reference files from chatgpt/references/ to the knowledge base / project files
3. Start asking for AWS architecture diagrams

This works with:

• ChatGPT → Custom GPT or Project with instructions + files
• Claude Projects → Project instructions + project knowledge
• Internal corporate AI agents → System prompt + knowledge base uploads
• Any LLM with file context → Paste prompt + reference content

Tip: Set temperature to 0.3 for consistent XML output.

/plugin marketplace add vidanov/aws-architecture-diagram-skill
/plugin install aws-architecture-diagram@vidanov-skills

Then use it:

/aws-architecture-diagram:aws-architecture-diagram

Or just ask naturally — the skill activates automatically when you mention AWS architecture diagrams.

mkdir -p ~/.claude/skills/aws-architecture-diagram
cp claude/SKILL.md ~/.claude/skills/aws-architecture-diagram/SKILL.md
cp -r references ~/.claude/skills/aws-architecture-diagram/references

# Global (all projects)
mkdir -p ~/.kiro/skills/aws-architecture-diagram
cp kiro/SKILL.md ~/.kiro/skills/aws-architecture-diagram/SKILL.md
cp -r references ~/.kiro/skills/aws-architecture-diagram/references

# Per-project
mkdir -p .kiro/skills/aws-architecture-diagram
cp kiro/SKILL.md .kiro/skills/aws-architecture-diagram/SKILL.md
cp -r references .kiro/skills/aws-architecture-diagram/references

Just ask to create an AWS architecture diagram:

Create an AWS architecture diagram for a serverless API with Lambda, DynamoDB, and API Gateway

Or with export:

Create an AWS architecture diagram as PNG for a real-time data pipeline with Kinesis, Lambda, and S3

"Create an event-driven order processing architecture with SQS, Lambda, DynamoDB, and EventBridge"

Download .drawio

"Create a real-time IoT analytics pipeline with Kinesis, Lambda, S3 data lake, and DynamoDB"

Download .drawio

"Create a 3-tier web application with CloudFront, ALB, ECS Fargate, Aurora, and ElastiCache"

Download .drawio

aws-architecture-diagram-skill/
├── README.md
├── LICENSE
├── chatgpt/
│   ├── PROMPT.md                       # Universal prompt (ChatGPT, Claude Projects, any AI)
│   └── references/                     # Same icons in .txt format for upload
│       ├── aws-icons-common.txt
│       ├── aws-icons-compute.txt
│       ├── aws-icons-database.txt
│       ├── aws-icons-integration.txt
│       ├── aws-icons-networking.txt
│       ├── aws-icons-storage.txt
│       ├── aws-icons-security.txt
│       ├── aws-icons-analytics-ml.txt
│       └── aws-icons-iot-migration-devtools.txt
├── kiro/
│   └── SKILL.md                        # Kiro CLI version
├── claude/
│   └── SKILL.md                        # Claude Code version
└── references/
    ├── aws-icons-compute.md            # Lambda, EC2, ECS, EKS, Fargate (25+ icons)
    ├── aws-icons-database.md           # DynamoDB, RDS, Aurora, ElastiCache (40+ icons)
    ├── aws-icons-integration.md        # API GW, SQS, SNS, EventBridge, Step Functions (35+ icons)
    ├── aws-icons-networking.md         # CloudFront, Route 53, VPC, ELB (40+ icons)
    ├── aws-icons-storage.md            # S3, EFS, EBS, Glacier, Backup (40+ icons)
    ├── aws-icons-security.md           # IAM, Cognito, KMS, WAF, Shield (45+ icons)
    ├── aws-icons-analytics-ml.md       # Kinesis, Athena, Bedrock, SageMaker (45+ icons)
    └── aws-icons-common.md             # Groups, general resources, edge styles, base template

draw.io AWS icons have two patterns with opposite strokeColor rules:

Confusing these patterns is the #1 cause of broken icons in AI-generated diagrams.

draw.io stencil names don't always match current AWS service names. Renamed services keep their old stencil identifiers:

Empty icon = wrong stencil name. The skill includes a verified icon catalog to prevent this.

The skill instructs the AI agent to:

1. Generate the .drawio XML
2. Export to PNG and visually review for broken icons or layout issues
3. Fix and re-export

This catches rendering problems that are invisible in raw XML.

These resIcon values do NOT render in draw.io — avoid them:

• mxgraph.aws4.dynamodb_table → use mxgraph.aws4.dynamodb instead
• mxgraph.aws4.dynamodb_stream → use mxgraph.aws4.dynamodb with label
• mxgraph.aws4.general_saml_token → use mxgraph.aws4.traditional_server

Why this approach wins:

• No runtime dependencies. It's a markdown file, not a server.
• Output is native .drawio XML you can open, edit, and version-control.
• Verified icon catalog documents which mxgraph.aws4.* names actually render vs. silently break.
• Opinionated layout: left-to-right flow with consistent spacing, not random placement.
• Works with any AI assistant: ChatGPT, Claude, Kiro CLI, or internal corporate agents.

Contributions are welcome! Here's how you can help:

1. Fork this repository
2. Create a branch for your feature or fix (git checkout -b my-feature)
3. Make your changes — add icons, fix stencil names, improve layout rules, etc.
4. Submit a Pull Request with a clear description of what you changed and why

You can also open an Issue if you:

• Found a broken icon or wrong stencil name
• Want to request support for a new AWS service
• Have suggestions for layout or styling improvements
• Encountered a bug with a specific AI agent

MIT