Manages multiple AI agent terminal sessions to orchestrate complex sub-tasks and parallelize development workflows.
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.
Key Features
01Seamless MCP server integration for web search and documentation tools
021 GitHub stars
03Real-time session status monitoring with visual indicators (running, waiting, idle, error)
04Flexible result retrieval with blocking and non-blocking modes
05Multi-session lifecycle management including start, stop, and restart
06Parent-child session hierarchy for advanced task delegation
Use Cases
01Orchestrating specialized agents equipped with different MCP tools for multi-step technical workflows
02Delegating long-running research tasks to a sub-agent while continuing to code in the main session
03Monitoring background AI processes and retrieving structured JSON output for automation
