Epic-specific acceptance + regression scenarios. The cross-epic manual-QA queue lives in ../../TEST_BACKLOG.md.
Each row corresponds to an exit criterion in the epic charter. Status reflects whether the acceptance scenario has been run + passed.
| ID | Scenario | Status |
|---|---|---|
| AT-01 | tinker capture "test note" returns in < 300ms; note retrievable in store |
not-tested |
| AT-02 | tinker capture (no arg) opens $EDITOR; empty save = no note; non-empty = stored |
not-tested |
| AT-03 | tinker recent returns last 20 notes for cwd ordered by timestamp desc |
not-tested |
| AT-04 | Install one-line works on macOS (brew), Linux (apt or cargo), Windows (choco or cargo) | partial |
| AT-05 | Full test suite (cargo test) green on all 3 OS in CI |
pass — BL-0001 |
Once an item closes with Test: pass, its acceptance scenario lands here as a regression to protect on every future change. Run before any v1.0-equivalent release.
| Area | Scenario | Last verified |
|---|---|---|
| CI baseline | cargo test green on macOS / Linux / Windows |
BL-0001 close |
Scenarios that can’t be cheaply automated and need human eyes — typically UX feel, install-path friction, error-message clarity.
| Scenario | Cadence | Notes |
|---|---|---|
| Time the install flow on a fresh machine; goal < 60s | Per minor release | First-success-time KPI |
tinker capture error messages on EACCES / no editor / disk full |
Per minor release | Error clarity check |
tinker recent output readable in 80-col + 200-col terminals |
Per minor release | Layout check |
AT-## (epic-scoped, monotonic).not-tested / pending / manual-verified / partial / pass / fail: <detail> / regression-needed / n/a.pass — BL-0001).deprecated if no longer relevant.