Close three critical review findings on the render-bounds-channel verdict

Verdict can no longer print a false EXACT on an on-grid end, no longer names a
bounds channel a content-derived render never consulted, and the grid-align doc
premise is corrected without implementing it.

src/shell/capture/CLAUDE.md

@@ -36,7 +36,14 @@ detail not covered there:
 - **`renderOffline` is the one seam both a fresh capture and a recipe replay
   cross**, which is why the refusal and both transient guards live there rather
   than in the action bodies — anything placed in `ResolveScopeSource` alone would
-  miss `RunRecaptureFromSource` entirely.
+  miss `RunRecaptureFromSource` entirely. The bounds channel is inside the backend
+  that seam calls, for the same reason: a replay must hand its window over exactly
+  the way a fresh capture does.
+- **The bounds channel is under live experiment**, and `capture.cpp`'s
+  `kBoundsChannel` is its single switch. On the time-selection channel the render
+  window travels in the project's own time selection, so `capture` snapshots and
+  restores that selection like any other state it borrows. Why there are two
+  channels: `src/core/capture/render_settings.h`'s `RenderBoundsChannel`.
 - **FX-bypass guard ordering.** `scope_resolve` reads the M10 provenance-assembly
   inputs (track/item selection, FX-chain identity) BEFORE the FX-bypass guard
   neutralizes the in-scope chain — provenance must see the chain as it really is,