poc

Summary

Quickly turn an idea into a testable proof of concept. Answer: Can this work? What is the smallest experiment to prove it? What evidence do we need before building the real thing?

Outputs: code sketch, experiment plan, test harness, architecture notes, and a go/iterate/stop recommendation.

Delegate feasibility analysis and POC design to a research-focused engineering agent.

When to use

• Idea is vague and needs validation.
• Comparing multiple approaches.
• Need a quick prototype, benchmark, or integration spike.
• Want a recommendation based on constraints (time, cost, platform, risk).

When not to use

• User needs production-ready code or full hardening.
• Problem is routine and well understood.
• Task requires broad refactoring, not a narrow experiment.