PROBLEM CLUE · CAPABILITY UPGRADES

Inconsistent results with no clear way to improve them

Model quality is rarely the whole story. Prompts, context, source quality, model choice, and task decomposition all influence the outcome.

CHECKLIST

Get these facts straight first.

  • Whether the current product is being used correctly
  • Which layer a prompt, skill, model, or plugin improves
  • Whether the task, inputs, versions, and outputs are reproducible

SOFTWARE MATCHES

GITHUB TREASURES