# awslabs/agent-plugins architecture

> Architecture diagram of the GitHub repository awslabs/agent-plugins, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2026-09-30). It is an AI-made overview: check details against the source.

- Interactive diagram: https://gitdiagram.com/awslabs/agent-plugins
- Repository: https://github.com/awslabs/agent-plugins

## Overview

Agent Plugins for AWS package reusable expertise for AI coding agents: skills guide workflows, MCP servers expose live services, hooks automate guardrails, and references supply supporting knowledge. The documented user journey is to install a plugin, invoke a relevant capability while architecting, deploying, or operating on AWS, and review the generated result. Sampled code shows concrete DSQL query-safety, deployment-diagram processing, and SageMaker data/model workflows. Most plugin workflows are documented but their implementation files are not present in the supplied tree; edges therefore reflect only README statements or visible calls.

Main entry points: [`plugins/databases-on-aws/skills/dsql/mcp/tools/safe_query.py`](https://github.com/awslabs/agent-plugins/blob/main/plugins/databases-on-aws/skills/dsql/mcp/tools/safe_query.py) (Safe SQL builder), [`plugins/sagemaker-ai/skills/dataset-transformation/code_templates/transformation.py`](https://github.com/awslabs/agent-plugins/blob/main/plugins/sagemaker-ai/skills/dataset-transformation/code_templates/transformation.py) (Dataset transformation), [`plugins/sagemaker-ai/skills/finetuning/code_templates/rlvr.py`](https://github.com/awslabs/agent-plugins/blob/main/plugins/sagemaker-ai/skills/finetuning/code_templates/rlvr.py) (Model fine-tuning), [`plugins/sagemaker-ai/skills/model-evaluation/code_templates/llmaaj_evaluator.py`](https://github.com/awslabs/agent-plugins/blob/main/plugins/sagemaker-ai/skills/model-evaluation/code_templates/llmaaj_evaluator.py) (Model evaluation)

## Diagram (Mermaid)

```mermaid
flowchart TD

subgraph group_packaging["Plugin capabilities"]
  node_plugin["AWS agent plugins"]
  node_skills["Guided skills"]
  node_mcp["MCP connections"]
  node_hooks["Hooks and guardrails"]
  node_references["Reference knowledge"]
end

subgraph group_deploy["Deployment diagrams"]
  node_diagram["Diagram post-processing"]
  node_diagram_checks["Diagram validation<br/>[validate_drawio.py]"]
  node_diagram_icons["Icon colors and badges<br/>[fix_icon_colors.py]"]
  node_diagram_badges["Step badges<br/>[fix_step_badges.py]"]
end

subgraph group_dsql["Aurora DSQL"]
  node_safe_sql["Safe SQL builder<br/>[safe_query.py]"]
  node_dsql_service["Aurora DSQL tools"]
end

subgraph group_sagemaker["SageMaker AI"]
  node_dataset_transform["Dataset transformation<br/>[transformation.py]"]
  node_transform_tools["Transformation helpers"]
  node_finetune["Model fine-tuning<br/>[rlvr.py]"]
  node_reward["RLVR reward function"]
  node_evaluation["Model evaluation"]
  node_deployment["Model deployment"]
  node_dataset_format["Dataset format detection<br/>[format_detector.py]"]
end

node_developer(("AI coding agent user"))

node_developer -->|"installs and invokes"| node_plugin
node_plugin -->|"packages"| node_skills
node_plugin -->|"packages"| node_mcp
node_plugin -->|"packages"| node_hooks
node_plugin -->|"packages"| node_references
node_skills -.->|"guides diagram work"| node_diagram
node_safe_sql -->|"builds SQL for"| node_dsql_service
node_dataset_transform -.->|"uses helpers"| node_transform_tools
node_finetune -->|"configures reward source"| node_reward
node_evaluation -.->|"evaluates then deploys"| node_deployment

click node_diagram "https://github.com/awslabs/agent-plugins/blob/main/plugins/deploy-on-aws/scripts/lib/post_process_drawio.py"
click node_diagram_checks "https://github.com/awslabs/agent-plugins/blob/main/plugins/deploy-on-aws/scripts/lib/validate_drawio.py"
click node_diagram_icons "https://github.com/awslabs/agent-plugins/blob/main/plugins/deploy-on-aws/scripts/lib/fix_icon_colors.py"
click node_diagram_badges "https://github.com/awslabs/agent-plugins/blob/main/plugins/deploy-on-aws/scripts/lib/fix_step_badges.py"
click node_safe_sql "https://github.com/awslabs/agent-plugins/blob/main/plugins/databases-on-aws/skills/dsql/mcp/tools/safe_query.py"
click node_dataset_transform "https://github.com/awslabs/agent-plugins/blob/main/plugins/sagemaker-ai/skills/dataset-transformation/code_templates/transformation.py"
click node_transform_tools "https://github.com/awslabs/agent-plugins/blob/main/plugins/sagemaker-ai/skills/dataset-transformation/scripts/transformation_tools.py"
click node_finetune "https://github.com/awslabs/agent-plugins/blob/main/plugins/sagemaker-ai/skills/finetuning/code_templates/rlvr.py"
click node_reward "https://github.com/awslabs/agent-plugins/blob/main/plugins/sagemaker-ai/skills/finetuning/templates/rlvr_reward_function_source_template.py"
click node_evaluation "https://github.com/awslabs/agent-plugins/blob/main/plugins/sagemaker-ai/skills/model-evaluation/code_templates/llmaaj_evaluator.py"
click node_deployment "https://github.com/awslabs/agent-plugins/blob/main/plugins/sagemaker-ai/skills/model-deployment/code_templates/deploy-oss-bedrock.py"
click node_dataset_format "https://github.com/awslabs/agent-plugins/blob/main/plugins/sagemaker-ai/skills/dataset-evaluation/scripts/format_detector.py"

classDef toneNeutral fill:#f8fafc,stroke:#334155,stroke-width:1.5px,color:#0f172a
classDef toneBlue fill:#dbeafe,stroke:#2563eb,stroke-width:1.5px,color:#172554
classDef toneAmber fill:#fef3c7,stroke:#d97706,stroke-width:1.5px,color:#78350f
classDef toneMint fill:#dcfce7,stroke:#16a34a,stroke-width:1.5px,color:#14532d
classDef toneRose fill:#ffe4e6,stroke:#e11d48,stroke-width:1.5px,color:#881337
classDef toneIndigo fill:#e0e7ff,stroke:#4f46e5,stroke-width:1.5px,color:#312e81
classDef toneTeal fill:#ccfbf1,stroke:#0f766e,stroke-width:1.5px,color:#134e4a
class node_plugin,node_skills,node_mcp,node_hooks,node_references,node_developer toneBlue
class node_diagram,node_diagram_checks,node_diagram_icons,node_diagram_badges toneAmber
class node_safe_sql,node_dsql_service toneMint
class node_dataset_transform,node_transform_tools,node_finetune,node_reward,node_evaluation,node_deployment,node_dataset_format toneRose
```

## Components

### Plugin capabilities

- **AWS agent plugins**
- **Guided skills**
- **MCP connections**
- **Hooks and guardrails**
- **Reference knowledge**

### Deployment diagrams

- [Diagram post-processing](https://github.com/awslabs/agent-plugins/blob/main/plugins/deploy-on-aws/scripts/lib/post_process_drawio.py): `plugins/deploy-on-aws/scripts/lib/post_process_drawio.py`
- [Diagram validation](https://github.com/awslabs/agent-plugins/blob/main/plugins/deploy-on-aws/scripts/lib/validate_drawio.py): `plugins/deploy-on-aws/scripts/lib/validate_drawio.py`
- [Icon colors and badges](https://github.com/awslabs/agent-plugins/blob/main/plugins/deploy-on-aws/scripts/lib/fix_icon_colors.py): `plugins/deploy-on-aws/scripts/lib/fix_icon_colors.py`
- [Step badges](https://github.com/awslabs/agent-plugins/blob/main/plugins/deploy-on-aws/scripts/lib/fix_step_badges.py): `plugins/deploy-on-aws/scripts/lib/fix_step_badges.py`

### Aurora DSQL

- [Safe SQL builder](https://github.com/awslabs/agent-plugins/blob/main/plugins/databases-on-aws/skills/dsql/mcp/tools/safe_query.py): `plugins/databases-on-aws/skills/dsql/mcp/tools/safe_query.py`
- **Aurora DSQL tools**

### SageMaker AI

- [Dataset transformation](https://github.com/awslabs/agent-plugins/blob/main/plugins/sagemaker-ai/skills/dataset-transformation/code_templates/transformation.py): `plugins/sagemaker-ai/skills/dataset-transformation/code_templates/transformation.py`
- [Transformation helpers](https://github.com/awslabs/agent-plugins/blob/main/plugins/sagemaker-ai/skills/dataset-transformation/scripts/transformation_tools.py): `plugins/sagemaker-ai/skills/dataset-transformation/scripts/transformation_tools.py`
- [Model fine-tuning](https://github.com/awslabs/agent-plugins/blob/main/plugins/sagemaker-ai/skills/finetuning/code_templates/rlvr.py): `plugins/sagemaker-ai/skills/finetuning/code_templates/rlvr.py`
- [RLVR reward function](https://github.com/awslabs/agent-plugins/blob/main/plugins/sagemaker-ai/skills/finetuning/templates/rlvr_reward_function_source_template.py): `plugins/sagemaker-ai/skills/finetuning/templates/rlvr_reward_function_source_template.py`
- [Model evaluation](https://github.com/awslabs/agent-plugins/blob/main/plugins/sagemaker-ai/skills/model-evaluation/code_templates/llmaaj_evaluator.py): `plugins/sagemaker-ai/skills/model-evaluation/code_templates/llmaaj_evaluator.py`
- [Model deployment](https://github.com/awslabs/agent-plugins/blob/main/plugins/sagemaker-ai/skills/model-deployment/code_templates/deploy-oss-bedrock.py): `plugins/sagemaker-ai/skills/model-deployment/code_templates/deploy-oss-bedrock.py`
- [Dataset format detection](https://github.com/awslabs/agent-plugins/blob/main/plugins/sagemaker-ai/skills/dataset-evaluation/scripts/format_detector.py): `plugins/sagemaker-ai/skills/dataset-evaluation/scripts/format_detector.py`

### Other components

- **AI coding agent user**

## Connections

- AI coding agent user → AWS agent plugins: installs and invokes (evidence: `README.md`)
- AWS agent plugins → Guided skills: packages (evidence: `README.md`)
- AWS agent plugins → MCP connections: packages (evidence: `README.md`)
- AWS agent plugins → Hooks and guardrails: packages (evidence: `README.md`)
- AWS agent plugins → Reference knowledge: packages (evidence: `README.md`)
- Guided skills → Diagram post-processing: guides diagram work
- Safe SQL builder → Aurora DSQL tools: builds SQL for (evidence: `plugins/databases-on-aws/skills/dsql/mcp/tools/safe_query.py`)
- Dataset transformation → Transformation helpers: uses helpers
- Model fine-tuning → RLVR reward function: configures reward source (evidence: `plugins/sagemaker-ai/skills/finetuning/code_templates/rlvr.py`)
- Model evaluation → Model deployment: evaluates then deploys