[BUG] Regression: ccd-cli v2.1.78 crashes with `invalid opcode` on non-AVX CPUs (same as fixed #19907)

Preflight Checklist

• [x] I have searched existing issues and this hasn't been reported yet
• [x] This is a single bug report (please file separate reports for different bugs)
• [x] I am using the latest version of Claude Code

What's Wrong?

The invalid opcode / SIGILL crash on non-AVX CPUs that was fixed after #19907 / #20019 has regressed in v2.1.78.

Claude Desktop deploys ccd-cli v2.1.78 to remote servers via SSH. On servers with QEMU Virtual CPU (only SSE/SSE2, no AVX), the binary crashes after ~10 seconds with trap invalid opcode. This is the same root cause as #19907 — the binary uses CPU instructions beyond baseline x86_64.

v2.1.72 worked perfectly on the same servers for days. The crash started immediately after Claude Desktop auto-updated and deployed v2.1.78.

Note: #19907 and #20019 are closed and locked, so I cannot comment there. This is a new issue to track the regression.

Server: Ubuntu 22.04, linux-x64
CPU: QEMU Virtual CPU version 2.5+ (flags: sse, sse2 only)
RAM: 2GB
Reproduced on 2 different servers, with both root and non-root users.

What Should Happen?

The ccd-cli binary should work on all x86_64 CPUs, including those without AVX/SSE4 support (baseline x86_64/SSE2). This worked correctly in v2.1.72.

Error Messages/Logs

# dmesg on remote server:
traps: ccd-cli[4888] trap invalid opcode ip:3263914 sp:7ffef4fb3d00 error:0 in ccd-cli[2d00000+3e3c000]

# CPU info:
model name : QEMU Virtual CPU version 2.5+
CPU flags: sse sse2 (no sse4, avx, avx2)

# ssh.log — working (v2.1.72):
2026-03-22 00:06:13 Spawn confirmed
2026-03-22 00:40:58 kill (user-initiated after 34 min of normal work)

# ssh.log — broken (v2.1.78):
2026-03-22 01:51:05 Spawn confirmed
2026-03-22 01:51:17 Exited, code=-1, duration=12667ms

Steps to Reproduce

1. Have a VPS with QEMU Virtual CPU (no AVX support), Ubuntu 22.04
2. Connect via Claude Desktop SSH remote
3. CLI v2.1.72 works fine — sessions run for 30-60+ minutes
4. Claude Desktop auto-updates, deploys ccd-cli v2.1.78 to the server
5. Every ccd-cli process now crashes after ~10 seconds with code=-1
6. dmesg shows: trap invalid opcode
7. UI shows: "RPC call to server.ping timed out"

Claude Model

Opus

Is this a regression?

Yes, this worked in a previous version

Last Working Version

2.1.72

Claude Code Version

2.1.78

Platform

Anthropic API

Operating System

Ubuntu/Debian Linux

Terminal/Shell

Other

Additional Information

This is the same class of bug as #19907 and #20019 (both closed and locked).
The fix that was applied after those issues appears to have been lost in v2.1.78.

Previously affected versions: 2.1.14, 2.1.15 (fixed sometime before 2.1.72)
Currently affected: 2.1.78
Root cause: likely a Bun runtime update that reintroduced AVX-dependent instructions

Timeline:

• Before 00:40 (Mar 22, 2026): CLI v2.1.72, everything works
• ~00:40: Claude Desktop restarts, deploys new claude-ssh binary + CLI v2.1.78
• After 00:40: every ccd-cli process crashes with code=-1 after ~10s