Begin with seed 104.
Candidate v0.9 under a stock timeout creates two drafts.
Demo workspace / Synthetic fixtures only
Inspect the run.
Keep the evidence.
Choose a condition and seed, review the resulting artifacts, and prepare a mock gate report. This workspace never calls external models, tools, or CI.
DEMO — SYNTHETIC DATA Simulation only
Harbor Parts / PO-204Replay workbench8 fixed runs / condition
Acceptance criterion Exactly one draft per orderBaseline met 8 / 8Candidate met 6 / 8
Seeded run matrix
| Seed | Baseline | Candidate | Inspect |
|---|---|---|---|
| 101 | 1 draft | 1 draft | |
| 102 | 1 draft | 1 draft | |
| 103 | 1 draft | 1 draft | |
| 104 | 1 draft | 2 drafts | |
| 105 | 1 draft | 1 draft | |
| 106 | 1 draft | 1 draft | |
| 107 | 1 draft | 1 draft | |
| 108 | 1 draft | 2 drafts |
candidate / seed 104
Criterion failedTwo drafts. One task.
The retry and late original response both reach draft creation. The candidate misses the exactly-one-draft criterion.
Purchase orderAPO-204-A
- Order reference
- PO-204
- Vendor
- Fictional Harbor Parts
- Origin
- Retry response
Expected draft
Purchase orderBPO-204-B
- Order reference
- PO-204
- Vendor
- Fictional Harbor Parts
- Origin
- Late original response
Duplicate artifact
Source: fixture/po-204.json · timeout · simulation only
Human adjudication / mock state
Keep the decision reviewable.
An engineer reviews the artifact and records why the criterion is met or failed.
Data dictionary & simulation boundary
PO-204 is fictional. Seeds 101–108 select fixed stub artifacts. Candidate timeout fails on 104 and 108; delay fails on 108; normal meets 8/8. Baseline meets 8/8. Drafts A and B denote artifact origins, not real purchase orders. Owner and review entries are local mock state. All external sends, billing, tool writes, and CI actions are absent.
Switch to baseline.
Baseline v0.8 suppresses the duplicate under the same fixture.
Record a rationale.
Review changes a local mock state; the report labels it explicitly.