massgen-develops-massgen

Use MassGen to test backend functionality (automation) or UI/UX (visual evaluation).

• Run MassGen in automation mode to test backend logic and agent coordination.
• Use visual evaluation to assess terminal UI/UX quality like colors and layout.
• Two mutually exclusive workflows for different improvement goals.
• Automation mode avoids visual noise with clean, parseable logs.

• Clean output with ~10 parseable lines vs 3,000+ ANSI codes.
• LOG_DIR printed on first line for log directory location.
• status.json provides real-time monitoring with progress and agent status.
• Exit codes: 0=success, 1=config, 2=execution, 3=timeout, 4=interrupted.
• Workspace isolation enables safe parallel execution.
• Multiple background monitors can track tokens, errors, progress, coordination.

1. 1.Run MassGen with --automation and specify config and question.
2. 2.Parse LOG_DIR from first line to locate log directory.
3. 3.Monitor status.json every 2 seconds for real-time updates.
4. 4.Read final answer from [log_dir]/final/[winner]/answer.txt after completion.
5. 5.Launch background monitor scripts (e.g., token, error, progress) to track metrics.