Does Agent Deck support background execution?

Yes, sessions run in the background by default, allowing you to trigger a task and retrieve the output later when needed.

Can I use MCP servers with Agent Deck sub-agents?

Yes, you can attach MCP servers like Exa or Firecrawl to specific sessions to give your sub-agents web search or documentation capabilities.

What is the main purpose of Agent Deck?

Agent Deck allows you to launch and manage 'sub-agents' or secondary Claude sessions from your main terminal to handle specialized tasks in parallel.

How do I know when a sub-agent has finished its task?

You can check the session status using the 'agent-deck status' command; a 'waiting' status usually indicates the agent is ready with a response.

Agent Deck

Name: Agent Deck
Author: artwist-polyakov

byartwist-polyakov

•

生产力与工作流

Manages multiple AI agent terminal sessions to orchestrate complex sub-tasks and parallelize development workflows.

关于

Agent Deck is a CLI-based session manager designed to empower Claude to spawn and control sub-agents. It allows users to launch background sessions for specialized tasks—such as web research, documentation analysis, or code refactoring—while maintaining focus in the primary terminal. With support for MCP servers and detailed status monitoring, it provides a robust framework for agent orchestration, making it ideal for developers who need to delegate complex, time-consuming tasks to secondary AI instances without losing their current context.

主要功能

Seamless MCP server integration for web search and documentation tools
1 GitHub stars
Real-time session status monitoring with visual indicators (running, waiting, idle, error)
Flexible result retrieval with blocking and non-blocking modes
Multi-session lifecycle management including start, stop, and restart
Parent-child session hierarchy for advanced task delegation

使用场景

Orchestrating specialized agents equipped with different MCP tools for multi-step technical workflows
Delegating long-running research tasks to a sub-agent while continuing to code in the main session
Monitoring background AI processes and retrieving structured JSON output for automation

关于

主要功能

Seamless MCP server integration for web search and documentation tools
1 GitHub stars
Real-time session status monitoring with visual indicators (running, waiting, idle, error)
Flexible result retrieval with blocking and non-blocking modes
Multi-session lifecycle management including start, stop, and restart
Parent-child session hierarchy for advanced task delegation

使用场景

Orchestrating specialized agents equipped with different MCP tools for multi-step technical workflows
Delegating long-running research tasks to a sub-agent while continuing to code in the main session
Monitoring background AI processes and retrieving structured JSON output for automation