AWS adds session memory and shell-command execution to AI agents
AWS has extended Bedrock AgentCore with two practical features for coding agents: persistent storage for session files and direct shell-command execution. An agent can now return to the same workspace after a pause, and commands such as npm test, pytest or git push run in the same container, without extra orchestration through the model. The platform also streams stdout and stderr in real time, which speeds up checking edits and debugging.
AI-processed from AWS Machine Learning Blog; edited by Hamidun News
AWS has updated its Bedrock AgentCore Runtime service: AI agents can now keep sessions across restarts and run shell commands directly from the model.
Why do AI agents need persistent storage?
Previously, an agent's session lived inside a separate micro-VM and lost all its files once it stopped. Now an agent can keep its working state across restarts, so saved dependencies or generated code, for example, are no longer lost.
Which commands can an AWS Bedrock AI agent run?
An agent can run predictable shell commands, such as tests or git operations, directly inside its own environment, without needing external tools.
Why should AI agents run shell commands?
To perform predictable operations such as tests, git commands and dependency installation safely inside the model's isolated environment.
- Oct 31, 2026Worldwide Express hands disputed invoices to AI agents and cuts unresolved disputes by 75%
- Oct 30, 2026AWS gives coding agents a live site preview straight from an isolated session
- Oct 30, 2026MoEngage Opens 400 Private APIs to AI Agents Through AgentCore Gateway, Without Building Its Own MCP Servers
- Oct 29, 2026NVIDIA Nemotron 3.5 Super VL: 120B Parameters, 12B Active, Running on AWS
Key Questions Answered
Why do AI agents need persistent storage?
Previously, an agent's session lived inside a separate micro-VM and lost all its files once it stopped. Now an agent can keep its working state across restarts, so saved dependencies or generated code, for example, are no longer lost.
Which commands can an AWS Bedrock AI agent run?
An agent can run predictable shell commands, such as tests or git operations, directly inside its own environment, without needing external tools.
Why should AI agents run shell commands?
To perform predictable operations such as tests, git commands and dependency installation safely inside the model's isolated environment.
Test Frontier LLMs & Autonomous Agents in One Workspace
Run Claude Opus 5.5, GPT-6, Gemini 3.1, and DeepSeek side-by-side in Nomad AI (nomadnet.ai) — unified model routing, deep research, and multimodal workflows under one subscription.
The AI world, distilled — once a week
Seven stories that actually mattered, hand-picked. No noise, no reposts, no press releases.
Done! Check your inbox for a confirmation.