Prompt Template
A prompt with variables, versioned like code.
A reusable prompt with placeholders filled at runtime. Treating templates as versioned artifacts — reviewed, diffed, rolled back — is basic hygiene once prompts drive production behaviour.
Prompts are configuration with the power of code. Without versioning, you cannot reproduce an incident or roll back a bad change.
Versioning prompts is the practice that makes everything else possible: it is what lets you attribute a behaviour change to a specific edit, roll back a bad one, and reproduce an incident. Without it, a quality regression has no diff to inspect.
Finding the change nobody deployed
Output quality shifted and no code was released.
Check the template history. A prompt edit is a behaviour change and needs the same review and the same rollout discipline as code. Teams that version prompts find this in minutes; teams that do not spend a week looking at the model vendor.
- docsPrompt engineering overview ↗
Ties each prompting technique to a success criterion and to empirical evaluation, rather than presenting them as a bag of tricks.
- docsEvaluation best practices ↗
Evals as structured tests: task-specific test sets, continuous evaluation, logging, and human calibration of automated scoring.
Prompt Versioning
Treating prompts as versioned, reviewable artifacts.
Prompt Engineering
The craft of writing instructions a model reliably follows.
Regression Test
A test that catches a specific past failure coming back.
Evaluation · Evals
How you know the system is actually working.