Understand this story
Beginning November 11, 2026, AWS rolled out an upgrade to both new and existing Amazon Bedrock AgentCore harnesses, adopting the open-source Strands harness and adding built-in tools, managed default configurations, and deterministic lifecycle hooks. The new version supports automatic prompt caching, proactive context management, task planning, subagent delegation, web search, and fine-grained file system access. Four types of lifecycle events can run allow-or-deny checks through AWS Lambda and can also send notifications to Amazon SNS or Amazon EventBridge. Existing users do not need to adopt new resource types or integrations, and existing explicit configurations remain unchanged. There are currently no other factually substantiated developments.
AI-generated from public reports
Latest development · Report
AWS Introduces… for the Amazon Bedrock AgentCore HarnessDevelopments
1 confirmed developments
- Source:AWS Machine Learning
Reports and discussion timeline
Browse reports and discussion by time. Their evidence levels remain distinct.
Nov 20
Unite.AI
AWS Introduces… for the Amazon Bedrock AgentCore Harness
AWS has upgraded the Amazon Bedrock AgentCore harness, adopting the open-source Strands harness and adding evolving managed defaults, built-in tools, and deterministic lifecycle hooks. The hooks can use Lambda to allow or deny actions before and after invocations and tool execution, with the same restrictions also covering sub-agent calls. The new experience began rolling out to both new and existing harnesses on November 11, 2026, requiring no customer action, while existing explicit configurations remain unchanged.
Unite.AIRelated discussion
Unite.AI joined this discussion
This post contributes to heat; it is not a formal report or confirmed fact.
AWS Machine LearningFirst-partySelected
Amazon Bedrock AgentCore Harness Upgrades Strands’ Default Capabilities and Lifecycle Hooks
The Amazon Bedrock AgentCore Harness upgrade adopts the open-source Strands harness and adds built-in tools, managed default configurations, and lifecycle hooks, without requiring existing users to adopt new resource types or integrations. The new version supports automatic prompt caching, proactive context management, task planning, subagent delegation, web search, and fine-grained file-system access. Context compaction can reduce token costs by approximately 34% compared with other leading harnesses. Four types of lifecycle events can perform allow-or-deny checks through AWS Lambda and can also send notifications to Amazon SNS or Amazon EventBridge.
AI Trends