AWS Secrets Manager released a secret safety skill within the aws-core plugin for the Agent Toolkit for AWS to securely manage secrets in AI coding agents’ workflows. This skill prevents raw secret values from appearing in the AI model’s context window or session logs by steering the agent to use the secret implicitly and resolving actual secret values only at execution in an isolated child process.
This two-layer approach ensures that sensitive information remains protected throughout agent interactions, enhancing security for developers using AI assistants. The skill supports multiple agents and is available in all regions where Secrets Manager operates. Verification of the protection effectiveness relies on how fully developers adopt the skill in their workflows.
Source & further reading
AWS Secrets Manager introduces safe secrets handling in the Agent Toolkit for AWS ↗Source published 2026-06-17. This report was prepared with AI assistance from the linked primary-source material. Claims and limitations should be checked against the original announcement or research.