This project demonstrates an "Infinite Agentic Loop" using a two-prompt system with Claude Code, enabling AI agents to collaboratively generate evolving content iterations based on specifications. It's designed for developers and researchers exploring advanced AI orchestration and content generation workflows.

How It Works

The system utilizes a custom Claude Code slash command (/project:infinite) to manage multiple AI agents. It analyzes a specification file, performs directory reconnaissance to understand existing iterations, and then deploys sub-agents with distinct creative directions. Wave management ensures coordinated generation, particularly in "infinite" mode, where agents produce successive, increasingly sophisticated iterations until context limits are reached.

Quick Start & Requirements

• Install: Copy .claude/commands/infinite.md to ~/.claude/commands/ for global use within Claude Code.
• Prerequisites: Requires Claude Code.
• Usage: Execute /project:infinite <spec_file> <output_dir> <count> within Claude Code.

Highlighted Details

• Supports single generation, small (5), large (20), and continuous "infinite" generation modes.
• Iterations are managed in waves for resource optimization.
• Agents are designed to ensure uniqueness and spec compliance.
• Can be extended to generate sets of files or integrated into an MCP server.

Maintenance & Community

• Project is experimental and associated with IndyDevDan's AI coding content.
• Further AI coding tips are available on the IndyDevDan YouTube channel.

Licensing & Compatibility

• No license is explicitly stated in the README.

Limitations & Caveats

The project is experimental and relies on the specific Claude Code environment. The README does not specify a license, which may impact commercial use or integration into closed-source projects.