Change Evaluator predicts how your existing test cases would respond to a change you've made — classifying each relevant case as PASS, FAIL, or UNCERTAIN — so you get an early read on impact before you push.
What it does
After you've made a change, Change Evaluator reasons about each relevant test case and predicts the outcome:
- PASS — the change is consistent with what the case expects.
- FAIL — the change conflicts with the case's expected behaviour; likely a regression to review.
- UNCERTAIN — the case is affected but the outcome can't be predicted with confidence; worth a human look.
It's a fast, early signal — a way to catch likely breakage and ambiguity before you spend a full test cycle finding them.
When to use it
- After implementing a change, before pushing or opening a pull request.
- To decide where to focus manual testing — start with the FAIL and UNCERTAIN cases.
- As a follow-up to Regression Preventer: that skill flags what's at risk before you build; Change Evaluator predicts the actual outcome once the change exists.
How to run it
With the Testmo skills installed (see the hub article), point it at your change and ask for the prediction. For example:
I've updated the password validation rules.
Evaluate our relevant test cases and tell me which would pass, fail, or are uncertain.The skill uses Semantic Search to find the affected cases, compares each against your change, and returns the verdicts with reasoning. It reads your cases and analyses your change locally, so a read-only connection is sufficient.
What you get back
A per-case verdict (PASS / FAIL / UNCERTAIN) with a short justification for each, so you can triage the likely failures and ambiguous cases immediately rather than waiting for a full test run to surface them.