This is a patterns guide for running multiple AI coding agents in parallel using dmux, a tmux-based pane manager. You create separate panes with `n`, each running its own agent session (Claude Code, Codex, whatever), then merge results back with `m`. The skill covers five solid patterns: research plus implementation in parallel, splitting multi-file features, test-and-fix loops, cross-harness work, and parallel code reviews. It also documents an ECC helper script that automates git worktree creation per agent so they don't step on each other's file changes. Useful when you have a chunky task that splits cleanly into independent pieces, though you'll burn tokens faster since each pane is a full session.
npx -y skills add affaan-m/ecc --skill dmux-workflows --agent claude-codeInstalls into .claude/skills of the current project.
Select a file.
juliusbrussee/caveman
mattpocock/skills
obra/superpowers
forrestchang/andrej-karpathy-skills
vercel-labs/skills