diff --git a/src/xrpld/app/consensus/ConsensusExtensionsDesign.md b/src/xrpld/app/consensus/ConsensusExtensionsDesign.md
index d32ee5f15..505a91e5a 100644
--- a/src/xrpld/app/consensus/ConsensusExtensionsDesign.md
+++ b/src/xrpld/app/consensus/ConsensusExtensionsDesign.md
@@ -486,12 +486,15 @@ would make a retained latch's accepted witness unreplayable.
 
 The destination account's SignerList remains the actual remote authority. Its
 keys, weights, optional operator signer, and rotation policy are client/operator
-configuration that source consensus cannot verify. The committee cap matches
-`STTx::maxMultiSigners()`; reserving a separate operator entry reduces the
-maximum validator roster usable by that destination account. A lower remote
-threshold weakens the claimed custody policy, while a higher threshold may
-strand an otherwise sufficient source witness. Target ledger validation and the
-returning XPOP are separate from this authorization decision.
+configuration that source consensus cannot verify. Export's 32-member source
+cap matches the expanded `STTx::maxMultiSigners()` capacity used by network-0
+signing rules. A destination without `featureExpandedSignerList` supports only
+eight entries and must configure a smaller committee; reserving a separate
+operator entry likewise reduces the maximum validator roster usable by that
+destination account. A lower remote threshold weakens the claimed custody
+policy, while a higher threshold may strand an otherwise sufficient source
+witness. Target ledger validation and the returning XPOP are separate from
+this authorization decision.
 
 The extended proposal machinery is enabled when either feature needs signed
 sidecar fields. Do not make Export depend on RNG availability just because RNG
diff --git a/src/xrpld/app/hook/detail/applyHook.cpp b/src/xrpld/app/hook/detail/applyHook.cpp
index b9884b5db..5f589ffef 100644
--- a/src/xrpld/app/hook/detail/applyHook.cpp
+++ b/src/xrpld/app/hook/detail/applyHook.cpp
@@ -4103,7 +4103,7 @@ readEntropySnapshot(ApplyView& view)
         sle->getFieldU16(sfEntropyDenominator)};
 }
 
-// byteCount must be a multiple of 32.
+// Callers normalize byteCount to a multiple of 32.
 // minTier is the CALLER'S stated class requirement (a required hook API
 // argument — there is deliberately no network-wide default). Count and
 // denominator policy is available separately through entropy_status().
@@ -4117,7 +4117,7 @@ fairRng(
     if (byteCount > 512)
         byteCount = 512;
 
-    // force the byte count to be a multiple of 32
+    // Preserve the caller's 32-byte block invariant defensively.
     byteCount &= ~0b11111;
 
     if (byteCount == 0)
diff --git a/src/xrpld/app/tx/detail/Import.cpp b/src/xrpld/app/tx/detail/Import.cpp
index d179c236f..1ecfdd214 100644
--- a/src/xrpld/app/tx/detail/Import.cpp
+++ b/src/xrpld/app/tx/detail/Import.cpp
@@ -943,8 +943,8 @@ Import::preflight(PreflightContext const& ctx)
         };
     if (hasInsufficientQuorum(quorum, validationCount))
     {
-        JLOG(ctx.j.warn()) << "Import: xpop did not contain an 80% quorum for "
-                              "the txn it purports to prove. "
+        JLOG(ctx.j.warn()) << "Import: xpop did not contain the required "
+                              "quorum for the txn it purports to prove. "
                            << tx.getTransactionID();
         return temMALFORMED;
     }
diff --git a/src/xrpld/consensus/ConsensusParms.h b/src/xrpld/consensus/ConsensusParms.h
index 21f63f212..55eb78cb5 100644
--- a/src/xrpld/consensus/ConsensusParms.h
+++ b/src/xrpld/consensus/ConsensusParms.h
@@ -104,13 +104,13 @@ struct ConsensusParms
      *  round (e.g. restarting after a crash) enters as proposing=false
      *  and cannot generate commitments until consensus promotes it to
      *  proposing — which takes at least one full round of observing.
-     *  Waiting the full 10s just delays the inevitable ZERO-entropy
+     *  Waiting the full 10s just delays the inevitable non-zero Tier-1
      *  fallback and slows recovery for the restarting node (it can't
      *  catch up until the survivors close a ledger).
      *
      *  3s is long enough for commits to propagate on any reasonable
      *  network, but short enough that a missing-node scenario recovers
-     *  quickly via the ZERO-entropy fallback path.
+     *  quickly via the Tier-1 fallback path.
      */
     std::chrono::milliseconds rngPIPELINE_TIMEOUT = std::chrono::seconds{3};