Here is a patch.

diff --git a/agent/divert-scd.c b/agent/divert-scd.c
index 1e5de4671..bb42dd3b4 100644
--- a/agent/divert-scd.c
+++ b/agent/divert-scd.c
@@ -517,6 +517,9 @@ agent_card_ecc_kem (ctrl_t ctrl, const unsigned char *ecc_ct,

Fixed with new GPGRT_PROCESS_STDIO_NUL flag.

Pushed the changes:

• Inheriting HANDLEs has been not working accurately
  • Before the fix, all HANDLEs were inherited
  • Only specified HANDLEs are inherited by: rE0b01950237ab: w32:spawn: Fix inheriting HANDLEs.
• HANDLEs by w32_open_null were leaked
  • fixed by: rEf28bf71b86ba: w32:spawn: Fix resource leaks for HANDLEs by w32_open_null.

T7723 fix by rE311fb769d1dd: w32:spawn: New flag GPGRT_PROCESS_STDIO_NUL.

Before implementing this feature, it's better to fix T7723: gpgrt:w32: Fix for inheriting stdin/stdout/stderr with "NUL", and do some clean up.

If we will fix gpgconf using GPGRT_PROCESS_STDIO_NUL, we will need to fix gpg-connect-agent to see if it's NUL or not.

Here is an experimental change to support the feature.

I'm testing the following patch with experimental change of libgpg-error.

In libgpg-error, we have: rE65114f24e13f: w32: More changes to the extended length path handling.

Fixed in 2.5.8.

secp256k1 failure:
https://lists.gnupg.org/pipermail/gnupg-users/2025-June/067731.html

Done by: rM8caa7cc517eb: Use sysconf as a fallback mechanism in the initialization.

OK. I'll add a code for setting the fallback value in _gpgme_io_subsystem_init and use it from get_max_fds.

For issues of get_max_fds, I created a sub task, although it seems not the direct cause of this particular problem.

I test following test program (gcc -o t-gmf t-gmf.c) on Debian machine of S390x.

In the log, we can observe duplicated lines generated by
https://dev.gnupg.org/source/gpgme/browse/master/src/posix-io.c$545
Example is like:

2025-05-19 20:16:35 gpgme[21970.55d7]   _gpgme_io_spawn: check: fd[0] = 0x1c -> 0x1
2025-05-19 20:16:35 gpgme[21970.55d7]   _gpgme_io_spawn: check: fd[0] = 0x1c -> 0x1

Done in 1.11.1.

Done in 1.11.1.

Done in 1.11.1.

Reading https://openssl-library.org/files/blog/Request_to_Extend_IETF_WGLC_for_PQ_Key_Specifications.pdf ,
seed (with "S") is included in the private-key.

The commit rC23543b6c1497: Add mldsa_compute_keygrip and let private-key include "p". works well for me.

To support Dilithium, we need to extend data handling of libgcrypt.
I propose following changes:

• internal flag of PUBKEY_FLAG_BYTE_STRING to ask opaque MPI for data to be signed/verified.
• The format of data as: (data(raw)[(flags no-prefix)](value ...)[(label ...)][(random-override ...)]): message, context, and random. Optional no-prefix flag to ask specific way of signing, controlling the internal, for Known Answer Tests (siggen).

My test coverage was not good (even if I daily use Curve25519 on Gnuk Token).
Your analysis is correct.