AI Amigos

Verified playbooks

Reuse what has been tested, not what merely sounds clever.

A public playbook records a real job, its baseline, the test method, evidence, result, limitations, and review history. Private project details never become public automatically.

The public library is being built from reviewed submissions. Start a project, complete an evaluation, and prepare a sanitized proof pack from the AI Practice Workspace.

Publication standard

  1. One distinct job to be done and a named audience.
  2. At least one baseline, one evaluation method, and one limitation.
  3. No secrets, personal records, unsupported high-stakes claims, or private outputs.
  4. Human editorial approval, visible version history, and CC BY-SA 4.0 attribution.

How the loop compounds

Visitors can fork an approved playbook, run it on their own task, report an outcome, and help the next version become more useful. A generated answer is disposable; a reviewed workflow with evidence can improve over time.