Skip to content

Dogfooding

If you won’t use it, why should anyone else?

Before shipping a skill, prompt, workflow, or tool to others, use it yourself on real work. Discover friction firsthand.

  • New skills or prompts before sharing with the team
  • Workflows before documenting them as standards
  • Tools before recommending to clients
  • Patterns before encoding in governance
  • Friction you didn’t anticipate
  • Edge cases the happy path missed
  • Assumptions that don’t hold
  • Steps that feel obvious to the author but aren’t
  1. Build the thing
  2. Use it on real work (not toy examples)
  3. Note every point of friction
  4. Fix or document before shipping
  5. Repeat until it feels effortless

Not “does it work?” but “would I choose to use this?”

If you find yourself avoiding your own tool, that’s signal.

  • Testing on synthetic examples only
  • Shipping before personal validation
  • Dogfooding once, then never again (things drift)
  • Exempting yourself from your own standards