AWS claude-platform documentation change
Summary
Rephrased API feature parity description, managed agents documentation, compliance disclaimer, and workspace access management instructions
Security assessment
Changes clarify existing functionality without addressing security vulnerabilities. IAM policy instructions for workspace access remain functionally identical to previous version.
Diff
diff --git a/claude-platform/latest/userguide/feature-support.md b/claude-platform/latest/userguide/feature-support.md index 0c8ecaa4e..007ac5091 100644 --- a//claude-platform/latest/userguide/feature-support.md +++ b//claude-platform/latest/userguide/feature-support.md @@ -11 +11 @@ Claude Managed AgentsComplianceHow workspace membership is modeledFeatures not c -Claude Platform on AWS uses the Anthropic Messages API directly, which means you get full Messages API feature parity with the first-party Claude API (except where noted in Features not currently available): +Claude Platform on AWS uses the Anthropic Messages API directly. You get full feature parity with the first-party Claude API, except where noted in Features not currently available: @@ -38 +38 @@ Claude Platform on AWS uses the Anthropic Messages API directly, which means you -Claude Managed Agents are available on Claude Platform on AWS. Agents, sessions, environments (cloud container configurations), credential vaults, and memory stores are first-class IAM resources; see [IAM actions](./iam-actions.html) for the action reference and [Using the Claude Console](./console.html) for the corresponding Claude Console pages. The Anthropic Agent SDK works against Claude Platform on AWS without a separate integration step — point it at the Claude Platform on AWS base URL and authenticate with SigV4 or an API key. +Claude Managed Agents are available on Claude Platform on AWS. Agents, sessions, environments, credential vaults, and memory stores are first-class IAM resources. See [IAM actions](./iam-actions.html) for the action reference and [Using the Claude Console](./console.html) for the corresponding pages. The Anthropic Agent SDK works with Claude Platform on AWS without additional integration. Point it at the Claude Platform on AWS base URL and authenticate with SigV4 or an API key. @@ -57 +57 @@ The following Claude Managed Agents capabilities are not currently available on -This service is a third-party offering provided by Anthropic and is not included within the scope of standard AWS compliance programs, certifications, or audit reports (such as SOC, ISO, or HIPAA eligibility). Customers are solely responsible for performing their own due diligence to ensure that this third-party offering meets their regulatory, legal, and compliance requirements. Before processing sensitive or regulated data, you should review Anthropic’s official compliance documentation via the [Anthropic Trust Center](https://trust.anthropic.com/). See the [AWS Service Terms](https://aws.amazon.com/service-terms/) for more information. +Anthropic provides this service as a third-party offering. It is not covered by standard AWS compliance programs, certifications, or audit reports (such as SOC, ISO, or HIPAA eligibility). Customers are solely responsible for performing their own due diligence to ensure that this third-party offering meets their regulatory, legal, and compliance requirements. Before processing sensitive or regulated data, you should review Anthropic’s official compliance documentation via the [Anthropic Trust Center](https://trust.anthropic.com/). See the [AWS Service Terms](https://aws.amazon.com/service-terms/) for more information. @@ -63 +63 @@ On the first-party Claude API, access is governed by users-to-workspaces members -This means workspace access is granted and revoked by modifying IAM policies — through AWS Organizations SCPs, permission boundaries, identity-attached policies, or resource-level conditions — not by managing per-workspace membership in a Claude Console UI. See [IAM policies](./iam-policies.html) for examples. +Grant and revoke workspace access by modifying IAM policies (SCPs, permission boundaries, identity-attached policies, or resource-level conditions) rather than managing per-workspace membership in the Claude Console. See [IAM policies](./iam-policies.html) for examples.