Amazon DocumentDB (with MongoDB compatibility) now available as a skill in the Agent Toolkit for AWS
AI agents can now provision, migrate, tune, and review DocumentDB clusters using 7 built-in expert workflows—with full IAM and CloudTrail guardrails.
View original announcement →Visual Summary
graph TD
A{{DocumentDB Agent Skill}}:::announced
B(AWS MCP Server):::compute
C(Amazon DocumentDB):::storage
D([Seven Workflows]):::feature
E([IAM Guardrails]):::feature
F((AI Coding Agent)):::external
G(AWS CloudTrail):::compute
H([Sandboxed Execution]):::feature
I(AWS CLI):::compute
F ==>|"invokes"| A
A ==>|"executes via"| B
B -->|"manages"| C
A -->|"follows"| D
B -->|"enforces"| E
B -->|"logs to"| G
B -->|"runs in"| H
A -.->|"standalone mode"| I
I -.->|"configures"| C
classDef announced fill:#ff9900,stroke:#ec7211,color:#fff,font-weight:bold
classDef compute fill:#e3f2fd,stroke:#1565c0,color:#1565c0
classDef storage fill:#e8f5e9,stroke:#2e7d32,color:#2e7d32
classDef feature fill:#fff3e0,stroke:#e65100,color:#e65100
classDef external fill:#f5f5f5,stroke:#616161,color:#616161
What's New
Amazon DocumentDB (with MongoDB compatibility) is now available as a specialized database skill within the Agent Toolkit for AWS, enabling AI coding agents to autonomously manage the full DocumentDB lifecycle. The skill encodes seven best-practice workflows—from cluster provisioning and schema design through migration, performance tuning, well-architected review, and major version upgrades—so agents can act on expert guidance without requiring developers to manually consult documentation. It is available at no additional charge and works both alongside the AWS MCP Server and as a standalone tool via the AWS CLI.
How It Works
- Seven structured workflows are bundled into the skill: cluster provisioning, schema design, MongoDB compatibility assessment, DMS-based migration with change data capture (CDC), performance tuning, a 41-check well-architected review, and major version upgrades—each providing step-by-step procedural guidance the agent follows deterministically.
- AWS MCP Server integration allows agents to execute AWS CLI commands and run diagnostic queries against live DocumentDB clusters through the Model Context Protocol (MCP), with all API calls routed through a single managed endpoint.
- IAM-based guardrails enforce access controls at the agent level, distinguishing agent actions from human actions and allowing organizations to set policies on what agents are permitted to do.
- CloudTrail audit logging captures every API call made by the agent, providing a full audit trail for compliance and incident investigation.
- Sandboxed Python execution lets agents run diagnostic scripts in an isolated environment, reducing the risk of unintended side effects on production infrastructure.
- Standalone AWS CLI mode (version 2.35.0+) allows teams without an MCP-compatible agent to invoke the skill locally, using
aws configure agent-toolkitto set up the environment. - Skill installation is handled via
npx skills add aws/agent-toolkit-for-aws/skillsor through native plugins for Claude Code, Codex, and Kiro, with the skill definition and reference scripts hosted publicly on GitHub.
Why It's Important
- Reduces operational errors by replacing ad-hoc agent reasoning about DocumentDB with tested, codified procedures, lowering the risk of misconfigured clusters or failed migrations.
- Accelerates developer velocity by eliminating the need to manually look up DocumentDB operational guidance; agents can complete complex multi-step tasks like DMS migrations with CDC in a single conversation.
- Bridges the training-cutoff gap by giving agents real-time, curated knowledge of DocumentDB best practices that may postdate the foundation model's training data.
- Enables safe agentic execution in enterprise environments through IAM guardrails and CloudTrail logging, addressing a key concern about unaudited agent activity on production AWS resources.
- Lowers the barrier to well-architected deployments by embedding a 41-check review directly into the agent workflow, making architectural best practices accessible without a dedicated review engagement.
- Supports migration use cases with a structured MongoDB compatibility assessment and DMS-based migration workflow, helping teams move existing MongoDB workloads to DocumentDB with confidence.
How It's Different
- Codified expertise vs. general model knowledge: Unlike relying on a foundation model's general AWS knowledge (which may be stale or incomplete), the skill provides up-to-date, AWS-authored procedural workflows that the agent follows step by step.
- 41-check well-architected review built in: No other publicly available DocumentDB tooling embeds a structured well-architected review at this level of granularity directly into an agent-executable workflow.
- Integrated security posture: Compared to agents executing AWS CLI commands directly in a local terminal, the MCP Server integration adds IAM-based access controls, CloudTrail logging, and sandboxed execution as first-class features rather than afterthoughts.
- Dual execution modes: The skill supports both MCP-connected agents (for live cluster interaction) and standalone AWS CLI execution, giving it broader compatibility than MCP-only solutions.
- Zero additional cost: The skill is free as part of the Agent Toolkit for AWS, unlike third-party database management or migration tools that typically carry licensing or usage fees.
- Agent-agnostic design: Works with Claude Code, Codex, Kiro, Cursor, and any MCP-compatible agent, avoiding lock-in to a single AI coding platform.
When to Prefer It
- Provisioning new DocumentDB clusters when you want an AI agent to configure instance types, replica sets, parameter groups, and VPC networking according to AWS best practices without manual intervention.
- Migrating from self-managed MongoDB when you need a structured compatibility assessment followed by a DMS-based migration with change data capture to minimize downtime.
- Performance troubleshooting when an agent should run diagnostic queries, analyze slow query logs, and recommend index or configuration changes on an existing cluster.
- Pre-production architectural reviews when you want to validate a DocumentDB deployment against a comprehensive 41-point well-architected checklist before go-live.
- Major version upgrades when you need a guided, step-by-step upgrade path that accounts for compatibility checks, pre-upgrade validation, and rollback considerations.
- Enterprise teams requiring auditability when agent actions on AWS infrastructure must be logged to CloudTrail and governed by IAM policies for compliance purposes.
- Teams preferring local execution when an MCP server is not available or desired, and the AWS CLI (v2.35.0+) is the preferred interface for running agent-driven workflows.
Availability
- General Availability (GA) as of July 13, 2026, with no preview or beta designation indicated.
- Global availability: The skill is a software artifact hosted on GitHub and executed locally or via the AWS MCP Server; it is not region-restricted, though the DocumentDB clusters it manages must reside in regions where Amazon DocumentDB is supported.
- Pricing: Available at no additional charge as part of the Agent Toolkit for AWS; standard Amazon DocumentDB, AWS DMS, and other AWS service costs apply when the skill provisions or interacts with those resources.
- Prerequisites: AWS CLI version 2.35.0 or later for standalone mode; IAM credentials configured locally;
uvinstalled for MCP proxy usage. - Compatible agents: Claude Code, Codex, Kiro, Cursor, and any MCP-compatible AI coding agent.
- Source and distribution: Skill definition, reference scripts, and documentation are publicly available on GitHub at
aws/agent-toolkit-for-awsunder the specialized database skills directory.
Related Resources
- https://aws.amazon.com/products/developer-tools/agent-toolkit-for-aws/
- https://docs.aws.amazon.com/agent-toolkit/latest/userguide/mcp-server.html
- https://github.com/aws/agent-toolkit-for-aws/tree/main/skills/specialized-skills/database-skills/amazon-documentdb
- https://docs.aws.amazon.com/agent-toolkit/latest/userguide/quick-start.html
- https://docs.aws.amazon.com/documentdb/latest/developerguide/
