diff --git a/ICSharpCode.Decompiler.Tests/Util/ResourcesFileTests.cs b/ICSharpCode.Decompiler.Tests/Util/ResourcesFileTests.cs
new file mode 100644
index 000000000..5c02a578e
--- /dev/null
+++ b/ICSharpCode.Decompiler.Tests/Util/ResourcesFileTests.cs
@@ -0,0 +1,253 @@
+// Copyright (c) 2026 Christoph Wille
+//
+// Permission is hereby granted, free of charge, to any person obtaining a copy of this
+// software and associated documentation files (the "Software"), to deal in the Software
+// without restriction, including without limitation the rights to use, copy, modify, merge,
+// publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons
+// to whom the Software is furnished to do so, subject to the following conditions:
+//
+// The above copyright notice and this permission notice shall be included in all copies or
+// substantial portions of the Software.
+//
+// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED,
+// INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR
+// PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE
+// FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+// OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+// DEALINGS IN THE SOFTWARE.
+
+using System;
+using System.Collections.Generic;
+using System.IO;
+using System.Linq;
+using System.Text;
+
+using ICSharpCode.Decompiler.Util;
+
+using NUnit.Framework;
+
+namespace ICSharpCode.Decompiler.Tests.Util
+{
+	/// <summary>
+	/// Every count and length in a .resources file is attacker-controlled once a crafted
+	/// assembly is opened. These tests pin down that each one is validated against the data
+	/// that actually exists before it drives an allocation, so a malformed file fails with a
+	/// BadImageFormatException instead of an out-of-memory condition.
+	/// </summary>
+	[TestFixture]
+	public class ResourcesFileTests
+	{
+		const string DefaultReaderType = "System.Resources.ResourceReader, mscorlib";
+		const string DeserializingReaderType = "System.Resources.Extensions.DeserializingResourceReader, System.Resources.Extensions, Version=4.0.0.0, Culture=neutral, PublicKeyToken=cc7b13ffcd2ddd51";
+
+		// ResourceTypeCode values as written into the data section.
+		const int TypeCodeString = 1;
+		const int TypeCodeByteArray = 0x20;
+		const int TypeCodeStream = 33;
+		const int TypeCodeStartOfUserTypes = 0x40;
+
+		[Test]
+		public void WellFormedFile_ReadsAllEntryKinds()
+		{
+			var blob = Build(DeserializingReaderType, new[] { "MyType, MyAssembly" }, new[] {
+				("text", StringPayload("hello")),
+				("blob", LengthPrefixedPayload(TypeCodeByteArray, new byte[] { 1, 2, 3 })),
+				("stream", LengthPrefixedPayload(TypeCodeStream, new byte[] { 4, 5 })),
+				("obj", SerializedObjectPayload(typeIndex: 0, kind: 2, new byte[] { 9, 8, 7 })),
+			});
+
+			var entries = new ResourcesFile(new MemoryStream(blob)).ToDictionary(e => e.Key, e => e.Value);
+
+			Assert.That(entries["text"], Is.EqualTo("hello"));
+			Assert.That(entries["blob"], Is.EqualTo(new byte[] { 1, 2, 3 }));
+			Assert.That(((MemoryStream)entries["stream"]!).ToArray(), Is.EqualTo(new byte[] { 4, 5 }));
+			var obj = (ResourceSerializedObject)entries["obj"]!;
+			Assert.That(obj.TypeName, Is.EqualTo("MyType, MyAssembly"));
+			using var objStream = obj.GetStream();
+			Assert.That(ReadAll(objStream), Is.EqualTo(new byte[] { 9, 8, 7 }));
+		}
+
+		[Test]
+		public void TypeCountBeyondStream_ThrowsInsteadOfAllocating()
+		{
+			// int.MaxValue string references is a multi-gigabyte allocation request; the
+			// header only has room for a handful of type names.
+			var blob = BuildHeaderOnly(numResources: 0, numTypes: int.MaxValue);
+
+			Assert.Throws<BadImageFormatException>(() => new ResourcesFile(new MemoryStream(blob)));
+		}
+
+		[Test]
+		public void ResourceCountBeyondStream_ThrowsBeforeReadingPositions()
+		{
+			// Each resource needs at least a 4-byte name hash and a 4-byte name position, so a
+			// count larger than the remaining bytes can be rejected up front rather than after
+			// the position array has been allocated and the read has run off the end.
+			var blob = BuildHeaderOnly(numResources: 100_000, numTypes: 0);
+
+			Assert.Throws<BadImageFormatException>(() => new ResourcesFile(new MemoryStream(blob)));
+		}
+
+		[Test]
+		public void ResourceNameLengthBeyondStream_ThrowsInsteadOfAllocating()
+		{
+			var blob = Build(DefaultReaderType, Array.Empty<string>(), new[] {
+				("name", StringPayload("value")),
+			}, forcedNameByteLength: int.MaxValue);
+
+			Assert.Throws<BadImageFormatException>(() => new ResourcesFile(new MemoryStream(blob)).ToList());
+		}
+
+		[TestCase(TypeCodeByteArray)]
+		[TestCase(TypeCodeStream)]
+		public void BinaryResourceLengthBeyondStream_ThrowsInsteadOfAllocating(int typeCode)
+		{
+			var payload = new MemoryStream();
+			using (var w = new BinaryWriter(payload, Encoding.UTF8, leaveOpen: true))
+			{
+				w.Write7BitEncodedInt(typeCode);
+				w.Write(int.MaxValue); // declared length, far beyond the file
+				w.Write(new byte[] { 1, 2, 3 });
+			}
+			var blob = Build(DefaultReaderType, Array.Empty<string>(), new[] { ("bin", payload.ToArray()) });
+
+			Assert.Throws<BadImageFormatException>(() => new ResourcesFile(new MemoryStream(blob)).ToList());
+		}
+
+		[Test]
+		public void SerializedObjectLengthBeyondStream_ThrowsInsteadOfAllocating()
+		{
+			var blob = Build(DeserializingReaderType, new[] { "MyType, MyAssembly" }, new[] {
+				("obj", SerializedObjectPayload(typeIndex: 0, kind: 1, new byte[] { 1 }, forcedLength: int.MaxValue)),
+			});
+			var obj = (ResourceSerializedObject)new ResourcesFile(new MemoryStream(blob)).Single().Value!;
+
+			Assert.Throws<BadImageFormatException>(() => obj.GetStream());
+		}
+
+		[Test]
+		public void SerializedObjectWithUnknownFormatKind_Throws()
+		{
+			// Only the four SerializationFormat kinds defined by System.Resources.Extensions
+			// are valid; anything else means the data section is not what the header claims.
+			var blob = Build(DeserializingReaderType, new[] { "MyType, MyAssembly" }, new[] {
+				("obj", SerializedObjectPayload(typeIndex: 0, kind: 99, new byte[] { 1 })),
+			});
+			var obj = (ResourceSerializedObject)new ResourcesFile(new MemoryStream(blob)).Single().Value!;
+
+			Assert.Throws<BadImageFormatException>(() => obj.GetStream());
+		}
+
+		static byte[] StringPayload(string value)
+		{
+			var ms = new MemoryStream();
+			using (var w = new BinaryWriter(ms, Encoding.UTF8, leaveOpen: true))
+			{
+				w.Write7BitEncodedInt(TypeCodeString);
+				w.Write(value);
+			}
+			return ms.ToArray();
+		}
+
+		static byte[] LengthPrefixedPayload(int typeCode, byte[] data)
+		{
+			var ms = new MemoryStream();
+			using (var w = new BinaryWriter(ms, Encoding.UTF8, leaveOpen: true))
+			{
+				w.Write7BitEncodedInt(typeCode);
+				w.Write(data.Length);
+				w.Write(data);
+			}
+			return ms.ToArray();
+		}
+
+		static byte[] SerializedObjectPayload(int typeIndex, int kind, byte[] data, int? forcedLength = null)
+		{
+			var ms = new MemoryStream();
+			using (var w = new BinaryWriter(ms, Encoding.UTF8, leaveOpen: true))
+			{
+				w.Write7BitEncodedInt(TypeCodeStartOfUserTypes + typeIndex);
+				w.Write7BitEncodedInt(kind);
+				w.Write7BitEncodedInt(forcedLength ?? data.Length);
+				w.Write(data);
+			}
+			return ms.ToArray();
+		}
+
+		static byte[] ReadAll(Stream stream)
+		{
+			var ms = new MemoryStream();
+			stream.CopyTo(ms);
+			return ms.ToArray();
+		}
+
+		/// <summary>
+		/// Writes the ResourceManager and RuntimeResourceSet headers up to and including the
+		/// resource and type counts, then stops. The stream ends where the type names would be.
+		/// </summary>
+		static byte[] BuildHeaderOnly(int numResources, int numTypes)
+		{
+			var ms = new MemoryStream();
+			using (var w = new BinaryWriter(ms, Encoding.UTF8, leaveOpen: true))
+			{
+				WriteHeader(w, DefaultReaderType, numResources, numTypes);
+			}
+			return ms.ToArray();
+		}
+
+		static void WriteHeader(BinaryWriter w, string readerType, int numResources, int numTypes)
+		{
+			w.Write(ResourcesFile.MagicNumber);
+			w.Write(1); // ResourceManager header version
+			w.Write(0); // bytes to skip (unused for header version 1)
+			w.Write(readerType);
+			w.Write("System.Resources.RuntimeResourceSet");
+			w.Write(2); // RuntimeResourceSet version
+			w.Write(numResources);
+			w.Write(numTypes);
+		}
+
+		/// <summary>
+		/// Builds a complete version-2 .resources file. The data of each entry is the raw
+		/// data section record (type code followed by payload).
+		/// </summary>
+		static byte[] Build(string readerType, string[] typeNames, IList<(string Name, byte[] Data)> entries, int? forcedNameByteLength = null)
+		{
+			var nameSection = new MemoryStream();
+			var dataSection = new MemoryStream();
+			var namePositions = new List<int>();
+			using (var nameWriter = new BinaryWriter(nameSection, Encoding.UTF8, leaveOpen: true))
+			{
+				foreach (var (name, data) in entries)
+				{
+					namePositions.Add((int)nameSection.Position);
+					byte[] nameBytes = Encoding.Unicode.GetBytes(name);
+					nameWriter.Write7BitEncodedInt(forcedNameByteLength ?? nameBytes.Length);
+					nameWriter.Write(nameBytes);
+					nameWriter.Write((int)dataSection.Position);
+					dataSection.Write(data);
+				}
+			}
+
+			var ms = new MemoryStream();
+			using (var w = new BinaryWriter(ms, Encoding.UTF8, leaveOpen: true))
+			{
+				WriteHeader(w, readerType, entries.Count, typeNames.Length);
+				foreach (var typeName in typeNames)
+					w.Write(typeName);
+				while (ms.Position % 8 != 0)
+					w.Write((byte)0);
+				foreach (var (name, _) in entries)
+					w.Write(name.GetHashCode()); // name hashes are not consulted by ResourcesFile
+				foreach (int position in namePositions)
+					w.Write(position);
+				// name section starts right after this data section offset field
+				int nameSectionStart = (int)ms.Position + sizeof(int);
+				w.Write(nameSectionStart + (int)nameSection.Length);
+				w.Write(nameSection.ToArray());
+				w.Write(dataSection.ToArray());
+			}
+			return ms.ToArray();
+		}
+	}
+}
diff --git a/ICSharpCode.Decompiler/Util/ResourcesFile.cs b/ICSharpCode.Decompiler/Util/ResourcesFile.cs
index 4034b2d5a..8227f94b3 100644
--- a/ICSharpCode.Decompiler/Util/ResourcesFile.cs
+++ b/ICSharpCode.Decompiler/Util/ResourcesFile.cs
@@ -152,18 +152,16 @@ namespace ICSharpCode.Decompiler.Util
 				throw new BadImageFormatException($"Unsupported resource set version: {version}");
 
 			numResources = reader.ReadInt32();
-			if (numResources < 0)
-			{
-				throw new BadImageFormatException(ResourcesHeaderCorrupted);
-			}
+			// Every resource contributes at least a 4-byte name hash and a 4-byte name position.
+			// Bounding the count by the bytes that remain also keeps numResources * 2 in
+			// GetStartPositions from overflowing.
+			CheckLength(numResources, bytesPerElement: 8);
 
 			// Read type positions into type positions array.
 			// But delay initialize the type table.
 			int numTypes = reader.ReadInt32();
-			if (numTypes < 0)
-			{
-				throw new BadImageFormatException(ResourcesHeaderCorrupted);
-			}
+			// Every type name is a length-prefixed string of at least one byte.
+			CheckLength(numTypes, bytesPerElement: 1);
 			typeTable = new string[numTypes];
 			for (int i = 0; i < numTypes; i++)
 			{
@@ -231,6 +229,22 @@ namespace ICSharpCode.Decompiler.Util
 			reader.Dispose();
 		}
 
+		/// <summary>
+		/// Validates a count or byte length read from the file before it is used to size an
+		/// allocation. Every element occupies at least <paramref name="bytesPerElement"/> bytes
+		/// in the stream, so a value that needs more bytes than remain after the current
+		/// position cannot describe real data; rejecting it up front keeps a crafted header from
+		/// forcing a multi-gigabyte allocation.
+		/// </summary>
+		void CheckLength(int count, int bytesPerElement)
+		{
+			long remaining = reader.BaseStream.Length - reader.BaseStream.Position;
+			if (count < 0 || (long)count * bytesPerElement > remaining)
+			{
+				throw new BadImageFormatException("Resources file corrupted: declared length exceeds the available data.");
+			}
+		}
+
 		public int ResourceCount => numResources;
 
 		public string GetResourceName(int index)
@@ -253,10 +267,7 @@ namespace ICSharpCode.Decompiler.Util
 				reader.Seek(pos, SeekOrigin.Begin);
 				// Can't use reader.ReadString, since it's using UTF-8!
 				int byteLen = reader.Read7BitEncodedInt();
-				if (byteLen < 0)
-				{
-					throw new BadImageFormatException("Resource name has negative length");
-				}
+				CheckLength(byteLen, bytesPerElement: 1);
 				bytes = new byte[byteLen];
 				// We must read byteLen bytes, or we have a corrupted file.
 				// Use a blocking read in case the stream doesn't give us back
@@ -448,20 +459,14 @@ namespace ICSharpCode.Decompiler.Util
 				case ResourceTypeCode.ByteArray:
 				{
 					int len = reader.ReadInt32();
-					if (len < 0)
-					{
-						throw new BadImageFormatException("Resource with negative length");
-					}
+					CheckLength(len, bytesPerElement: 1);
 					return reader.ReadBytes(len);
 				}
 
 				case ResourceTypeCode.Stream:
 				{
 					int len = reader.ReadInt32();
-					if (len < 0)
-					{
-						throw new BadImageFormatException("Resource with negative length");
-					}
+					CheckLength(len, bytesPerElement: 1);
 					byte[] bytes = reader.ReadBytes(len);
 					return new MemoryStream(bytes, writable: false);
 				}
@@ -548,8 +553,12 @@ namespace ICSharpCode.Decompiler.Util
 				if (usesSerializationFormat)
 				{
 					int kind = reader.Read7BitEncodedInt();
-					Debug.Assert(Enum.IsDefined(typeof(SerializationFormat), kind));
+					if (!Enum.IsDefined(typeof(SerializationFormat), kind))
+					{
+						throw new BadImageFormatException("Resources file corrupted: unknown serialization format.");
+					}
 					len = reader.Read7BitEncodedInt();
+					CheckLength(len, bytesPerElement: 1);
 				}
 				return reader.ReadBytes(len);
 			}