dmux-workflows

To run parallel AI agents across tools, this Skill uses dmux to orchestrate multiple agent sessions simultaneously. It creates parallel agent panes across Claude Code, Codex, and OpenCode, enabling split tasks to execute concurrently with git worktree isolation.

Dmux is a tmux-based pane manager used for multi-agent workflow orchestration. It creates parallel agent panes, distributes prompts to each pane, and handles pane merging to coordinate complex, split tasks across multiple AI coding tools.

Yes, you can coordinate Claude Code and Codex sessions using dmux. The Skill supports cross-tool synergy by managing multiple agent panes within a single dmux session for parallel execution and task splitting.

To start a dmux session for parallel agent workflows, initiate dmux and press 'n' to create new agent panes. Type prompts for each pane to assign specific tasks, and the session orchestrates parallel execution across your tools.

Yes, multi-agent orchestration supports git worktree isolation for parallel tasks. The Skill implements worktree isolation to ensure concurrent agent sessions in separate panes remain independent until results are merged.

Use parallel AI agent sessions when managing complex tasks that require splitting into multiple concurrent subtasks. If a workflow needs simultaneous execution across different tools like Claude Code and Codex, dmux coordinates the panes and merges the results.