The mlperf-common (primus-mllog) maintainers and AMD take the security of
this project seriously. Thank you for helping keep it and its users safe.
Please do not report security vulnerabilities through public GitHub issues, pull requests, or discussions.
Instead, use one of the private channels below:
-
GitHub Private Vulnerability Reporting (preferred). Go to the repository's Security tab and click "Report a vulnerability" to open a private advisory with the maintainers. See GitHub's guide on privately reporting a security vulnerability.
-
AMD Product Security (PSIRT). For AMD-wide product security concerns, you may also contact AMD PSIRT at psirt@amd.com. See the AMD Product Security page for more information.
To help us triage quickly, please provide as much of the following as possible:
- A description of the vulnerability and its potential impact.
- Steps to reproduce or a proof of concept.
- Affected version(s), commit SHA, and environment (ROCm version, GPU, PyTorch / Megatron / Primus versions).
- Any suggested mitigation, if known.
- We will acknowledge receipt of your report within 5 business days.
- We will provide an initial assessment and expected next steps, and keep you informed of progress toward a fix.
- We will coordinate the timing of any public disclosure with you and credit reporters who wish to be acknowledged.
Please give us a reasonable opportunity to address the issue before any public disclosure.
This project is distributed primarily as a source package installed from the
main branch or a pinned commit. Security fixes are applied to the latest
main; we recommend always tracking the most recent commit.
| Version | Supported |
|---|---|
main (latest) |
✅ |
| older commits | ❌ |