Validate pair seed simulation behavior
This commit is contained in:
@@ -26,6 +26,8 @@ The map audit currently sees:
|
||||
- Route simulation family coverage, so representative generated rows exercise
|
||||
every registered action and position family except documented special cases
|
||||
covered by dedicated smoke fixtures.
|
||||
- Pair seed simulation, so Insta/OF soft/hard metadata and formatter outputs
|
||||
prove locked determinism and pose-only reroll behavior.
|
||||
|
||||
## Architectural Finding
|
||||
|
||||
@@ -643,6 +645,8 @@ Near-term:
|
||||
scene/camera/clothing fields.
|
||||
- Keep same-room pair continuity synchronized in both assembled prompt text and
|
||||
`hardcore_row.scene_text`; `tools/prompt_smoke.py` covers this drift case.
|
||||
- Keep pair seed behavior synchronized across soft/hard rows; the route
|
||||
simulator now checks locked pair determinism and pose-only hard-action rerolls.
|
||||
|
||||
Medium-term:
|
||||
|
||||
|
||||
Reference in New Issue
Block a user