Repository navigation
Closed
Labels
bugSomething is broken or behaving incorrectly.Something is broken or behaving incorrectly.needs-triageIssue needs maintainer review and initial categorization.Issue needs maintainer review and initial categorization.
Description
Before submitting
- I searched existing issues and did not find a duplicate.
- I included enough detail to reproduce or investigate the problem.
Area
apps/server
Steps to reproduce
1. Run T3 Code on a machine whose CPU lacks AVX/AVX2 instruction sets (e.g., Intel Pentium Gold, Celeron, or legacy x86_64).
2. Configure WSL backend (or native host).
3. Navigate to Providers -> Antigravity.
4. Click to Install / Activate Antigravity provider.
5. Wait for the 682 MB download and archive extraction to complete.
Expected behavior
Either:
1. Pre-flight check: T3 Code checks CPU flags (avx/avx2) before downloading 682 MB and warns that Google Antigravity binary requires AVX instructions.
2. Clear validation error: If runtime validation fails due to SIGILL (signal 4 / exit code 132), UI clearly informs the user that the process crashed due to an illegal CPU instruction caused by missing AVX
support.
Actual behavior
T3 Code downloads the entire 682 MB archive, extracts it, attempts validation (`agy_acp_server.par --uid=`), crashes immediately on `SIGILL`, and surfaces only a generic message in UI:
> "The downloaded Antigravity runtime could not start in this environment."
Users assume archive corruption, file permission errors, or WSL communication glitches, repeatedly re-downloading ~682 MB.
Impact
Blocks work completely
Version or commit
v0.0.28
Environment
- OS: Windows 10/11 x64 with WSL2 backend (Ubuntu 22.04 LTS, kernel 6.18.33.2-microsoft-standard-WSL2) - CPU: Intel(R) Pentium(R) Gold G5400 @ 3.70GHz (flags: SSE4.1, SSE4.2; lacks AVX and AVX2) - Chipset: Intel B360
Logs or stack traces
From `server.trace.ndjson`:
{
"type": "effect-span",
"name": "AntigravityInstallation.validate",
"exit": {
"_tag": "Failure",
"cause": "AntigravityInstallationError: The downloaded Antigravity runtime could not start in this environment.\n [cause]: AcpTransportError: ACP transport operation read-process-exit-status failed.\n [cause]:
PlatformError: Unknown: ChildProcess.exitCode (.../agy_acp_server.par --uid=)\n [cause]: Error: Process interrupted due to receipt of signal: 'SIGILL'"
}
}
Kernel output (dmesg):
agy_acp_server.par: potentially unexpected fatal signal 4.
WSL (CaptureCrash): Capturing crash for pid: 1877, executable: .../agy_acp_server.par, signal: 4Screenshots, recordings, or supporting files
No response
Workaround
None...
• Upgrade CPU to an AVX/AVX2 capable processor (e.g. Intel Core i3/i5/i7 8th/9th Gen on LGA1151 or AMD Ryzen).
• Run the runtime binary under Intel SDE emulator (sde64), though with significant performance penalty.
• Use alternative providers in T3 Code that do not require Google's AVX runtime binary.
Metadata
Metadata
Assignees
Labels
bugSomething is broken or behaving incorrectly.Something is broken or behaving incorrectly.needs-triageIssue needs maintainer review and initial categorization.Issue needs maintainer review and initial categorization.
Activity
CMP-TIC commentedon Oct 31, 2026
I cannot try it anymore as i updated my cpu to an AVX2 capable. But when i did the problem continued.
But then i updated t3code, and now the problem is gone and i can see, connect and use antigravity.
meinarus commentedon Nov 5, 2026
Recently, ACP was added in the Antigravity CLI. My mini PC's CPU also doesn't have AVX. I tried installing the Antigravity CLI and it works just fine, while downloading the Antigravity ACP runtime also fails like yours. I wonder if they will migrate to using the ACP of the AGY CLI, since currently they are using some special workaround for antigravity haha.