diff --git a/scratchpad/SESSION_HANDOFF_20260803.md b/scratchpad/SESSION_HANDOFF_20260803.md
index 9b56c90..ed7b71d 100644
--- a/scratchpad/SESSION_HANDOFF_20260803.md
+++ b/scratchpad/SESSION_HANDOFF_20260803.md
@@ -307,3 +307,17 @@ MS-FIREFLY/santo=Conn Man (Steam DOWN -- winsock fallback, note!),
 GAMERSLAB/rajel=RajelAran, SCREECH-PC/Dave=Oracle -- confirm from
 session headers at scan time.  NEXT: log scan (crash @8:45PM ET Oracle
 lobby launch; [ghost] lines vs skating windows; #131 forensics).
+
+### #131 FALSE LOCK: FIXED -- miss-means-miss (2026-08-05)
+Root cause: sphere + AABB fallbacks answered where the 1995 card would
+MISS; exposed by LEVEL boresight over the SHORT blackhawk (ray clears the
+mesh, threads the spheres -- ring lit above its head; operator watched it
+live).  Fix: pick returns hit/TRUE-MISS/no-tree; spheres only for
+unreadable meshes (none exist); AABB only pre-tree.  Receipts: level
+sweep 0 locks (was 168 sphere answers); down-pitch 100% tri-sourced;
+walk matrix tri=18874/sphereFB=0, 156 hits across 16 zones on bhk1.
+New instruments: BT_LOCK_SWEEP / BT_LOCK_ENVELOPE / [picksrc] counters.
+BUILD RITUAL (stale-link flake, bit 3x today): rm build/Release/btl4.exe
+before the build, string-verify the new exe before deploying.
+FIELD NOTE: locking strictly tighter now; if era testers want linger,
+that becomes a sourced investigation (Draco's memory, #131 comment).