Agent Standup is a new open-source tool that turns your Claude Code session into a live team meeting. The lead agent gives orders, sub-agents join, share their screen, and report back. It's described as a bit silly, but also weirdly useful.
The tool works wherever you run Claude Code: VS Code, Cursor, a plain terminal, or the desktop app. Nothing needs to be configured, and nothing is sent anywhere. It reads the session files that Claude Code already writes, so it works everywhere Claude Code runs.
Agent Standup shows you who is waiting on whom and who is blocked, at zero token cost. You can see a live call that follows your Claude Code session with agents sharing their screen, showing real commands and real output, and the exact changes they make.
The tool is available as a VS Code extension, a standalone desktop app for macOS, or via npx for any terminal. It's free, open source, and works on Mac, Windows, and Linux. The project is still experimental, with support for other agent tools like Codex CLI and Gemini CLI on the roadmap.
The project addresses a growing need in the AI development community for better visibility into multi-agent workflows. As developers increasingly rely on AI agents for complex tasks, understanding what those agents are doing and how they coordinate becomes critical for debugging and optimization.