AWS announced flexible namespace variables for Amazon Bedrock AgentCore Memory on 28 August 2026. Developers can now organise long-term agent memories by dimensions such as organisation, tenant, team or environment. Our assessment: this is useful infrastructure for multi-tenant agents, but a namespace is only a security boundary when trusted identity and tested authorisation determine its value.
What happened
According to AWS, developers can define keys on a memory resource, reference those keys in a strategy namespace template and provide values through the CreateEvent API. AgentCore then substitutes the values during long-term memory extraction. A resource can define up to five keys, and the same key can be referenced by multiple strategies.
AWS says the capability is available at no additional cost in every region where AgentCore Memory is generally available. The announcement describes how memories can be organised and accessed; it does not say that adding a tenant-shaped key automatically validates who supplied it.
Why it matters
Long-term memory turns a one-off model interaction into persistent application state. In a shared service, a wrong namespace can therefore expose one customer’s context to another, mix production and test data, or give an agent an inaccurate history. That makes namespace selection part of the application’s authorisation model, not a formatting convenience.
The feature can reduce duplicated memory strategies for complex organisational structures. It also makes the intended boundary visible enough to document and test. The important design question is not simply whether the namespace contains a tenant ID, but whether that value comes from authenticated server-side context and is checked consistently across writes, retrieval and deletion.
The practical implication
IT Web Builder recommends defining a short namespace contract before rollout. Record what each key means, which identity claim supplies it, who may access it and how it changes when a person moves teams or an environment is retired. Avoid accepting tenant or organisation values directly from model output or untrusted client input.
Build negative tests as well as successful ones. Attempt cross-tenant reads and writes, missing-key requests, altered identifiers and replayed events. Confirm that logs show the resolved boundary without exposing the stored memory itself, and include deletion and retention behaviour in the test plan.
Start with a non-sensitive workload and a small number of clearly owned dimensions. More namespace keys can express a richer hierarchy, but they also create more combinations to reason about. These are IT Web Builder rollout recommendations, not a claim that flexible namespaces alone provide tenant isolation or compliance.
Key takeaways
Derive namespace values from authenticated server-side context.
Test denied cross-tenant access across writes, reads and deletion.
Document ownership, retention and lifecycle rules for every key.
Sources
This article is original analysis written by IT Web Builder. The reporting and product details below were used as references and are credited to their publishers.
AWS — Bedrock AgentCore Memory adds flexible namespace variablesPublished 28 August 2026. Feature limits and availability are attributed to AWS; the governance and testing recommendations are IT Web Builder Editorial analysis.↗IT Web Builder publishes original summaries and practical commentary. We do not reproduce source articles. Product claims and announcements are linked to their original publishers so readers can verify the details.