If you believe you have found a security vulnerability in mlir-aie, please report it privately rather than opening a public issue.
The preferred channel is GitHub's private vulnerability reporting:
https://github.com/Xilinx/mlir-aie/security/advisories/new
This opens a private advisory thread visible only to the maintainers. We will acknowledge reports as soon as we are able.
Please include, where possible:
- A description of the vulnerability and its potential impact.
- Steps to reproduce, or a minimal proof-of-concept.
- The commit hash or release tag the issue was observed against.
- Any suggested mitigation.
mlir-aie tracks the tip of main. Security fixes are applied to main
as soon as feasible; prior releases are not patched in place.
In-scope: code in this repository, including build tooling, Python bindings, and CI workflows.
Out-of-scope: vulnerabilities in upstream LLVM/MLIR (please report to the LLVM project) or in third-party dependencies (please report to the respective project).