{{CANONICAL}}
← Back to Tech News

Amazon Bedrock AgentCore Runtime introduces interactive shells for terminal access into agent sessions

Amazon Web Services has launched a new interactive shell capability for Amazon Bedrock AgentCore Runtime, enabling developers to access persistent terminal sessions directly within running agent environments. The new InvokeAgentRuntimeCommandShell API provides a WebSocket-based connection to a PTY-backed terminal inside isolated microVMs, complete with standard terminal features including color support, tab completion, keyboard shortcuts, and automatic reconnection after network interruptions. The enhancement specifically targets developers building and deploying coding agents such as Claude Code, OpenAI Codex, and Amazon Kiro on the AgentCore platform. Unlike the existing one-shot command execution API, the interactive shell maintains persistent state across commands within the same session, preserving environment variables, working directory context, and command history. Each session supports up to 10 concurrent shells identified by runtime session and shell IDs, allowing developers to operate multiple terminals simultaneously against the same or different microVMs for parallel development workflows. Developers can initiate interactive sessions using the AgentCore CLI with the command 'agentcore exec --it --runtime <runtime-arn>' and can resume disconnected sessions by providing the same session identifiers. The feature addresses a key gap in agent development workflows by providing direct access to inspect files, execute ad-hoc commands, and debug environment states within the containerized agent runtime.

Why It Matters

This release represents a significant improvement in the developer experience for AI agent development on AWS, bridging the gap between cloud-hosted AI services and traditional local development workflows. By providing direct terminal access to agent runtime environments, AWS is addressing a critical pain point for developers who need to debug, inspect, and iteratively develop coding agents. The persistent session capability and multi-shell support could accelerate adoption of cloud-based agent development platforms by offering the familiar terminal interface that developers expect, while maintaining the isolation and scalability benefits of microVM architecture.

Read Original Release →
Note

This summary is generated using AI analysis of the original press release. Always refer to the original source for complete details.