scenario-data-fixlisted
Install: claude install-skill YevheniiaQA/qa-agent-toolkit
# Scenario Data Fix
Scenarios written with invented data compile into tests that fail for reasons
unrelated to the feature. This skill closes the gap between "a scenario that
reads well" and "a scenario that can actually run".
## Two placeholder categories
`scenarios-generator` emits two kinds, and they are resolved differently.
**`{{TEST_DATA: key}}`** — static data that should already exist. This skill
resolves it by lookup, following the order below, and asks the engineer for
whatever remains.
**`{{GENERATED: key}}`** — values created fresh on each run. **These are not
yours to resolve.** Do not look them up, do not substitute a fixed value, and do
not ask the engineer for one. Verify instead that each carries a complete
generation rule — `rule`, `scope`, `reuse`, and `cleanup` where it creates data —
and pass it through untouched for the automation stage to implement as a fixture.
A `{{GENERATED: ...}}` placeholder that has been replaced with a concrete value is
a bug: it turns a per-run unique value into a shared constant, and the suite will
pass serially and collide under parallel workers.
If a placeholder uses neither notation, that is an upstream gap. Report it rather
than inventing a third convention.
## Resolution order
Work through sources in this order and stop at the first that answers:
1. **Global context** — `01-docs/context/test-data.yaml` and
`environments.yaml`. Existing test users, tenants, tokens, seed records.
2. **Derivable from contracts**