Bring me the problem.

A plain-language description of what it should help someone accomplish.

Idea, architecture, prototype, active system—or somewhere in between.

The assumptions, trade-offs, handoffs, risks, or conflicting directions you cannot resolve cleanly.

Name the decision, who owns it, and why it matters now.

For example: OpenClaw, LangGraph, a custom service, scripts, manual coordination, not sure, or nothing yet. This is separate from the AI model or coding assistant.

For example: send messages, change records, approve work, move money, place orders, or call an external service. Choose “Not sure” if you cannot tell.

This is a starting point, not a package recommendation.

Choose “Another agent-system problem” for research, operations, personal AI, monitoring, or another consequential agent system when you are unsure where it fits.