Coordinating multiple coding agents with nothing but tmux and codex
viksit · x · 2026-09-09
A minimal multi-agent setup requiring no orchestration framework: run codex in a tmux session as a coordinator, have it spawn background agents in new tmux windows (8-char names), each on its own git worktree and branch, and clean up windows when done. Agents communicate by natural-language instructions, e.g. telling tmux to change the UI color and tail a new log file.
More from coding & agent
- Speculative decoding explains why coding agents type fast on boilerplate but slow on new logic — RunAI_Coder · 2026-09-09
- Ben Lorica: practical self-improvement arrives in the AI harness long before RSI — bigdata · 2026-09-09
- ax: open-source CLI audits any site for AI coding agents via npx — EdenEmarco177 · 2026-09-09
- Cursor details harness that let thousands of agents make most commits to a browser for a week — KyeGomezB · 2026-09-09
- Dev dissects GPT-6 Astra's computer use loop, makes it 2.5x faster with Stagehand — TheMoonMidas · 2026-09-09
- Preregistered study finds both writing skill and CS background predict vibe-coding performance — dair_ai · 2026-09-09