diff --git a/t/lib-chunk.sh b/t/lib-chunk.sh
new file mode 100644
index 0000000000..a7cd9c3c6d
--- /dev/null
+++ b/t/lib-chunk.sh
@@ -0,0 +1,17 @@
+# Shell library for working with "chunk" files (commit-graph, midx, etc).
+
+# corrupt_chunk_file <fn> <chunk> <offset> <bytes>
+#
+# Corrupt a chunk-based file (like a commit-graph) by overwriting the bytes
+# found in the chunk specified by the 4-byte <chunk> identifier. If <offset> is
+# "clear", replace the chunk entirely. Otherwise, overwrite data <offset> bytes
+# into the chunk.
+#
+# The <bytes> are interpreted as pairs of hex digits (so "000000FE" would be
+# big-endian 254).
+corrupt_chunk_file () {
+	fn=$1; shift
+	perl "$TEST_DIRECTORY"/lib-chunk/corrupt-chunk-file.pl \
+		"$@" <"$fn" >"$fn.tmp" &&
+	mv "$fn.tmp" "$fn"
+}
diff --git a/t/lib-chunk/corrupt-chunk-file.pl b/t/lib-chunk/corrupt-chunk-file.pl
new file mode 100644
index 0000000000..cd6d386fef
--- /dev/null
+++ b/t/lib-chunk/corrupt-chunk-file.pl
@@ -0,0 +1,66 @@
+#!/usr/bin/perl
+
+my ($chunk, $seek, $bytes) = @ARGV;
+$bytes =~ s/../chr(hex($&))/ge;
+
+binmode STDIN;
+binmode STDOUT;
+
+# A few helpers to read bytes, or read and copy them to the
+# output.
+sub get {
+	my $n = shift;
+	return unless $n;
+	read(STDIN, my $buf, $n)
+		or die "read error or eof: $!\n";
+	return $buf;
+}
+sub copy {
+	my $buf = get(@_);
+	print $buf;
+	return $buf;
+}
+
+# read until we find table-of-contents entry for chunk;
+# note that we cheat a bit by assuming 4-byte alignment and
+# that no ToC entry will accidentally look like a header.
+#
+# If we don't find the entry, copy() will hit EOF and exit
+# (which should cause the caller to fail the test).
+while (copy(4) ne $chunk) { }
+my $offset = unpack("Q>", copy(8));
+
+# In clear mode, our length will change. So figure out
+# the length by comparing to the offset of the next chunk, and
+# then adjust that offset (and all subsequent) ones.
+my $len;
+if ($seek eq "clear") {
+	my $id;
+	do {
+		$id = copy(4);
+		my $next = unpack("Q>", get(8));
+		if (!defined $len) {
+			$len = $next - $offset;
+		}
+		print pack("Q>", $next - $len + length($bytes));
+	} while (unpack("N", $id));
+}
+
+# and now copy up to our existing chunk data
+copy($offset - tell(STDIN));
+if ($seek eq "clear") {
+	# if clearing, skip past existing data
+	get($len);
+} else {
+	# otherwise, copy up to the requested offset,
+	# and skip past the overwritten bytes
+	copy($seek);
+	get(length($bytes));
+}
+
+# now write out the requested bytes, along
+# with any other remaining data
+print $bytes;
+while (read(STDIN, my $buf, 4096)) {
+	print $buf;
+}