Reference — generated from the toolkit

Source of truth: toolkit/prompts/test/tests-from-acs.md. Edit it there; this page is regenerated on build.

Tests from acceptance criteria

Purpose: generate honest, behaviour-asserting tests from a feature's ACs.

Prompt

Generate tests for the attached acceptance criteria. Rules: assert observable BEHAVIOUR, not the implementation. No tautologies, no snapshot-everything. Do not test the mocks. Cover every positive, negative and boundary example in the ACs. For insurance date/currency logic, include timezone and rounding boundaries. If an AC is ambiguous, list it as a question rather than guessing an assertion. Do NOT modify any existing test.

Usage notes

• Output is reviewed against the AI-generated-test review standard; mutation score is the arbiter.

Failure modes

• Tests written to pass, not to verify (mitigated: behaviour/oracle instruction + mutation gate).
• Silently weakening existing tests (mitigated: "do not modify existing tests"; human approval required anyway).