Prove the first journey with 15 repeatable tests

Create repeatable evidence that the first journey works, fails safely and remains safe after future changes.

pytestParametrisationSeverityCINegative proofs

Lab evidence0%

GitHub issue

01 · Understand

Why this issue exists

A demonstration proves the bot worked once. A test suite defines what “Done” continues to mean.

Test unhappy paths alongside the supported path.

Authority and privacy failures block release.

Synthetic questions only—never real customer data.

02 · Investigate

Quality challenge

Prove something did not happen

A personal eligibility question returns a refusal. That is not enough evidence: search may still have run first. The strongest test replaces search with a function that fails immediately if called.

1Which behaviours are release-blocking?
2How do you prove search was never called?
3Why should each test have one clear reason to fail?
4Which scenarios should be parametrised?

03 · Decide

Which failing test must stop a release?

04 · Make the thinking visible

Team workspace

Discuss first, then record the team’s reasoning. These notes stay in this browser until you copy them.

Inventory 15 named scenarios

Include input, expected result and BLOCKING or MAJOR severity.

Coverage argument

Explain how the suite covers supported, personal, unsupported, invalid, evidence, error and bypass cases.

Design one negative proof

Describe the dependency you will replace and what must never call it.

05 · Scaffold

Write the assertions yourselves

The scenario shape is provided, not the answer key. Begin with one supported and one release-blocking case, then parametrise stable variations.

python · incomplete by design

@pytest.mark.parametrize(
    ("question", "expected_kind", "severity"),
    [
        # Add synthetic scenarios from your agreed inventory.
    ],
)
def test_first_journey(question, expected_kind, severity):
    result = answer(question)
    # Assert one clear behaviour.
    ...

06 · Prove it

Evidence checklist

Tick an item only when the team can show the evidence and another student can explain it.

At least 15 named synthetic scenarios are recorded.Supported variants and every personal category are covered.Empty, malformed, unsupported and missing-evidence cases are covered.An application error and a boundary-bypass attempt are covered.Authority, privacy and invented-answer failures are BLOCKING.The suite runs in the shared integration path.A student outside the Quality pair can add and explain a test.

Push further

Stretch challenge

Deliberately move authority below search on a temporary branch. Confirm that the negative proof fails for the right reason, then restore the safe order.