diff --git a/crates/core_simd/src/elements/float.rs b/crates/core_simd/src/elements/float.rs
index 5a628f2121e..67e4454e5e1 100644
--- a/crates/core_simd/src/elements/float.rs
+++ b/crates/core_simd/src/elements/float.rs
@@ -202,17 +202,20 @@ macro_rules! impl_trait {
             #[inline]
             fn to_bits(self) -> Simd<$bits_ty, LANES> {
                 assert_eq!(core::mem::size_of::<Self>(), core::mem::size_of::<Self::Bits>());
+                // Safety: transmuting between vector types is safe
                 unsafe { core::mem::transmute_copy(&self) }
             }
 
             #[inline]
             fn from_bits(bits: Simd<$bits_ty, LANES>) -> Self {
                 assert_eq!(core::mem::size_of::<Self>(), core::mem::size_of::<Self::Bits>());
+                // Safety: transmuting between vector types is safe
                 unsafe { core::mem::transmute_copy(&bits) }
             }
 
             #[inline]
             fn abs(self) -> Self {
+                // Safety: `self` is a float vector
                 unsafe { intrinsics::simd_fabs(self) }
             }
 
@@ -283,11 +286,13 @@ macro_rules! impl_trait {
 
             #[inline]
             fn simd_min(self, other: Self) -> Self {
+                // Safety: `self` and `other` are float vectors
                 unsafe { intrinsics::simd_fmin(self, other) }
             }
 
             #[inline]
             fn simd_max(self, other: Self) -> Self {
+                // Safety: `self` and `other` are floating point vectors
                 unsafe { intrinsics::simd_fmax(self, other) }
             }
 
diff --git a/crates/core_simd/src/lib.rs b/crates/core_simd/src/lib.rs
index 2632073622e..715f258f617 100644
--- a/crates/core_simd/src/lib.rs
+++ b/crates/core_simd/src/lib.rs
@@ -12,7 +12,7 @@
 #![cfg_attr(feature = "generic_const_exprs", feature(generic_const_exprs))]
 #![cfg_attr(feature = "generic_const_exprs", allow(incomplete_features))]
 #![warn(missing_docs)]
-#![deny(unsafe_op_in_unsafe_fn)]
+#![deny(unsafe_op_in_unsafe_fn, clippy::undocumented_unsafe_blocks)]
 #![unstable(feature = "portable_simd", issue = "86656")]
 //! Portable SIMD module.
 
diff --git a/crates/core_simd/src/masks.rs b/crates/core_simd/src/masks.rs
index 11d7288eccb..c36c336d8a2 100644
--- a/crates/core_simd/src/masks.rs
+++ b/crates/core_simd/src/masks.rs
@@ -68,6 +68,7 @@ macro_rules! impl_element {
             const FALSE: Self = 0;
         }
 
+        // Safety: this is a valid mask element type
         unsafe impl MaskElement for $ty {}
     }
 }
diff --git a/crates/core_simd/src/masks/to_bitmask.rs b/crates/core_simd/src/masks/to_bitmask.rs
index 954f88ea511..65d3ce9be65 100644
--- a/crates/core_simd/src/masks/to_bitmask.rs
+++ b/crates/core_simd/src/masks/to_bitmask.rs
@@ -16,11 +16,7 @@ where
 /// Converts masks to and from integer bitmasks.
 ///
 /// Each bit of the bitmask corresponds to a mask lane, starting with the LSB.
-///
-/// # Safety
-/// This trait is `unsafe` and sealed, since the `BitMask` type must match the number of lanes in
-/// the mask.
-pub unsafe trait ToBitMask: Sealed {
+pub trait ToBitMask: Sealed {
     /// The integer bitmask type.
     type BitMask;
 
@@ -34,12 +30,8 @@ pub unsafe trait ToBitMask: Sealed {
 /// Converts masks to and from byte array bitmasks.
 ///
 /// Each bit of the bitmask corresponds to a mask lane, starting with the LSB of the first byte.
-///
-/// # Safety
-/// This trait is `unsafe` and sealed, since the `BYTES` value must match the number of lanes in
-/// the mask.
 #[cfg(feature = "generic_const_exprs")]
-pub unsafe trait ToBitMaskArray: Sealed {
+pub trait ToBitMaskArray: Sealed {
     /// The length of the bitmask array.
     const BYTES: usize;
 
@@ -51,9 +43,9 @@ pub unsafe trait ToBitMaskArray: Sealed {
 }
 
 macro_rules! impl_integer_intrinsic {
-    { $(unsafe impl ToBitMask<BitMask=$int:ty> for Mask<_, $lanes:literal>)* } => {
+    { $(impl ToBitMask<BitMask=$int:ty> for Mask<_, $lanes:literal>)* } => {
         $(
-        unsafe impl<T: MaskElement> ToBitMask for Mask<T, $lanes> {
+        impl<T: MaskElement> ToBitMask for Mask<T, $lanes> {
             type BitMask = $int;
 
             fn to_bitmask(self) -> $int {
@@ -69,13 +61,13 @@ macro_rules! impl_integer_intrinsic {
 }
 
 impl_integer_intrinsic! {
-    unsafe impl ToBitMask<BitMask=u8> for Mask<_, 1>
-    unsafe impl ToBitMask<BitMask=u8> for Mask<_, 2>
-    unsafe impl ToBitMask<BitMask=u8> for Mask<_, 4>
-    unsafe impl ToBitMask<BitMask=u8> for Mask<_, 8>
-    unsafe impl ToBitMask<BitMask=u16> for Mask<_, 16>
-    unsafe impl ToBitMask<BitMask=u32> for Mask<_, 32>
-    unsafe impl ToBitMask<BitMask=u64> for Mask<_, 64>
+    impl ToBitMask<BitMask=u8> for Mask<_, 1>
+    impl ToBitMask<BitMask=u8> for Mask<_, 2>
+    impl ToBitMask<BitMask=u8> for Mask<_, 4>
+    impl ToBitMask<BitMask=u8> for Mask<_, 8>
+    impl ToBitMask<BitMask=u16> for Mask<_, 16>
+    impl ToBitMask<BitMask=u32> for Mask<_, 32>
+    impl ToBitMask<BitMask=u64> for Mask<_, 64>
 }
 
 /// Returns the minimum numnber of bytes in a bitmask with `lanes` lanes.
@@ -85,7 +77,7 @@ pub const fn bitmask_len(lanes: usize) -> usize {
 }
 
 #[cfg(feature = "generic_const_exprs")]
-unsafe impl<T: MaskElement, const LANES: usize> ToBitMaskArray for Mask<T, LANES>
+impl<T: MaskElement, const LANES: usize> ToBitMaskArray for Mask<T, LANES>
 where
     LaneCount<LANES>: SupportedLaneCount,
 {
diff --git a/crates/core_simd/src/ops.rs b/crates/core_simd/src/ops.rs
index 8dbae346b5f..5a077a469d8 100644
--- a/crates/core_simd/src/ops.rs
+++ b/crates/core_simd/src/ops.rs
@@ -33,6 +33,7 @@ where
 
 macro_rules! unsafe_base {
     ($lhs:ident, $rhs:ident, {$simd_call:ident}, $($_:tt)*) => {
+        // Safety: $lhs and $rhs are vectors
         unsafe { $crate::simd::intrinsics::$simd_call($lhs, $rhs) }
     };
 }
@@ -49,6 +50,7 @@ macro_rules! unsafe_base {
 macro_rules! wrap_bitshift {
     ($lhs:ident, $rhs:ident, {$simd_call:ident}, $int:ident) => {
         #[allow(clippy::suspicious_arithmetic_impl)]
+        // Safety: $lhs and the bitand result are vectors
         unsafe {
             $crate::simd::intrinsics::$simd_call(
                 $lhs,
@@ -91,6 +93,7 @@ macro_rules! int_divrem_guard {
                 // Nice base case to make it easy to const-fold away the other branch.
                 $rhs
             };
+            // Safety: $lhs and rhs are vectors
             unsafe { $crate::simd::intrinsics::$simd_call($lhs, rhs) }
         }
     };
diff --git a/crates/core_simd/src/ops/unary.rs b/crates/core_simd/src/ops/unary.rs
index 4ebea560fc6..4ad02215034 100644
--- a/crates/core_simd/src/ops/unary.rs
+++ b/crates/core_simd/src/ops/unary.rs
@@ -14,6 +14,7 @@ macro_rules! neg {
             #[inline]
             #[must_use = "operator returns a new vector without mutating the input"]
             fn neg(self) -> Self::Output {
+                // Safety: `self` is a signed vector
                 unsafe { intrinsics::simd_neg(self) }
             }
         })*
diff --git a/crates/core_simd/src/round.rs b/crates/core_simd/src/round.rs
index 556bc2cc1fe..e111f3e0494 100644
--- a/crates/core_simd/src/round.rs
+++ b/crates/core_simd/src/round.rs
@@ -30,6 +30,8 @@ macro_rules! implement {
                 $type: FloatToInt<I>,
                 I: SimdElement,
             {
+                // Safety: `self` is a vector, and `FloatToInt` ensures the type can be casted to
+                // an integer.
                 unsafe { intrinsics::simd_cast(self) }
             }
         }
diff --git a/crates/core_simd/src/vector.rs b/crates/core_simd/src/vector.rs
index 9eb51f957d1..fac7dca51f4 100644
--- a/crates/core_simd/src/vector.rs
+++ b/crates/core_simd/src/vector.rs
@@ -213,7 +213,7 @@ where
     #[inline]
     #[cfg(not(bootstrap))]
     pub fn cast<U: SimdElement>(self) -> Simd<U, LANES> {
-        // Safety: The input argument is a vector of a known SIMD type.
+        // Safety: The input argument is a vector of a valid SIMD element type.
         unsafe { intrinsics::simd_as(self) }
     }
 
@@ -624,61 +624,85 @@ pub unsafe trait SimdElement: Sealed + Copy {
 }
 
 impl Sealed for u8 {}
+
+// Safety: u8 is a valid SIMD element type, and is supported by this API
 unsafe impl SimdElement for u8 {
     type Mask = i8;
 }
 
 impl Sealed for u16 {}
+
+// Safety: u16 is a valid SIMD element type, and is supported by this API
 unsafe impl SimdElement for u16 {
     type Mask = i16;
 }
 
 impl Sealed for u32 {}
+
+// Safety: u32 is a valid SIMD element type, and is supported by this API
 unsafe impl SimdElement for u32 {
     type Mask = i32;
 }
 
 impl Sealed for u64 {}
+
+// Safety: u64 is a valid SIMD element type, and is supported by this API
 unsafe impl SimdElement for u64 {
     type Mask = i64;
 }
 
 impl Sealed for usize {}
+
+// Safety: usize is a valid SIMD element type, and is supported by this API
 unsafe impl SimdElement for usize {
     type Mask = isize;
 }
 
 impl Sealed for i8 {}
+
+// Safety: i8 is a valid SIMD element type, and is supported by this API
 unsafe impl SimdElement for i8 {
     type Mask = i8;
 }
 
 impl Sealed for i16 {}
+
+// Safety: i16 is a valid SIMD element type, and is supported by this API
 unsafe impl SimdElement for i16 {
     type Mask = i16;
 }
 
 impl Sealed for i32 {}
+
+// Safety: i32 is a valid SIMD element type, and is supported by this API
 unsafe impl SimdElement for i32 {
     type Mask = i32;
 }
 
 impl Sealed for i64 {}
+
+// Safety: i64 is a valid SIMD element type, and is supported by this API
 unsafe impl SimdElement for i64 {
     type Mask = i64;
 }
 
 impl Sealed for isize {}
+
+// Safety: isize is a valid SIMD element type, and is supported by this API
 unsafe impl SimdElement for isize {
     type Mask = isize;
 }
 
 impl Sealed for f32 {}
+
+// Safety: f32 is a valid SIMD element type, and is supported by this API
 unsafe impl SimdElement for f32 {
     type Mask = i32;
 }
 
 impl Sealed for f64 {}
+
+// Safety: f64 is a valid SIMD element type, and is supported by this API
 unsafe impl SimdElement for f64 {
     type Mask = i64;
 }