# Amazon — Amazon Bedrock AgentCore harness adds Strands harness defaults, built-in tools, and lifecycle hooks

- Company: Amazon (amazon.com)
- Announced: 2026-11-19T20:01:06+00:00
- Category: capability-change
- Coverage: not counted
- Announcement: yes
- Group: announcements
- Source: https://aws.amazon.com/blogs/machine-learning/amazon-bedrock-agentcore-harness-adds-strands-harness-defaults-built-in-tools-and-lifecycle-hooks/
- Record: https://forck.live/items/38852-amazon-bedrock-agentcore-harness-adds-strands-harness-defaults-built-in-tools
- Subject: Bedrock / Nova
- License: Apache 2.0

Amazon is upgrading the Amazon Bedrock AgentCore harness to run on Strands harness, its open-source agent harness released earlier this year under an Apache 2.0 license, and is adding lifecycle hooks for deterministic checks inside the agent loop. Agents that use the managed defaults now get a recommended model, a task-oriented system prompt, automatic prompt caching for Anthropic models on Amazon Bedrock, and proactive context management. Amazon says keeping context lean this way helps the harness run at about 34% lower token cost than other leading harnesses. New built-in tools cover planning, web retrieval, granular filesystem access, context offloading and task delegation, with the subagent tool sharing invocation limits and the customer's tool-call lifecycle hooks. Amazon says there is no new resource type or integration to adopt, and both updates arrive automatically in the AgentCore harness customers already use. Explicit configuration still takes precedence: a model ID, a replacement systemPrompt, a truncation policy or a fixed allowedTools list overrides the managed defaults, and Amazon warns that applications with a fixed security boundary should list only the tools they need because the recommended set can gain capabilities over time.

## Evidence

Verbatim from https://aws.amazon.com/blogs/machine-learning/amazon-bedrock-agentcore-harness-adds-strands-harness-defaults-built-in-tools-and-lifecycle-hooks/:

> When building an AI agent, selecting a model is only the first step. The agent still needs tools, instructions for finishing the job, a way to manage context growth in long sessions, and guardrails before it executes any consequential actions, like issuing a refund.
> 
> We are upgrading Amazon Bedrock AgentCore harness to use Strands harness, our open-source agent harness, released earlier this year, available under an Apache 2.0 license. AgentCore harness also introduces lifecycle hooks, so you can apply your own deterministic checks and notifications at defined points in the agent loop.
> 
> With this upgrade, agents that rely on the managed defaults get a recommended model, a task-oriented system prompt, automatic prompt caching (initially for Anthropic models on Amazon Bedrock), and proactive context management. They also get additional built-in tools for planning, web retrieval, filesystem access, context offloading, and task delegation. This gives agents greater capabilities during long tasks, allowing them to keep track of subtasks, reduce context window usage, and delegate work.

## Follow-ups

- Unite.AI (reported, 19 Nov): AWS Adds Strands Defaults and Lifecycle Hooks to AgentCore Harness — https://forck.live/items/38886-aws-adds-strands-defaults-and-lifecycle-hooks-to-agentcore-harness

---

Record: https://forck.live/items/38852-amazon-bedrock-agentcore-harness-adds-strands-harness-defaults-built-in-tools
Catalogue: https://forck.live/llms.txt
Current issue: https://forck.live/feed.md
