Skip to content

Security: underpass-ai/AXLR

SECURITY.md

Security policy

This policy covers every public Underpass AI repository that does not publish its own SECURITY.md. KMP and MADE publish their own; follow those for them.

Reporting a vulnerability

Do not open a public issue for an undisclosed vulnerability.

Use GitHub private vulnerability reporting on the affected repository: its Security tab, Report a vulnerability (https://github.com/underpass-ai/<repository>/security/advisories/new). If that form is unavailable, contact a maintainer through GitHub without publishing exploit details.

Please include:

  • the affected repository and release or commit
  • a minimal, redacted reproduction
  • the impact
  • any known mitigation

Do not publish credentials or an exploit before coordinated triage.

Supported versions

The projects are pre-1.0. Fixes land on main and ship in the next release where a repository publishes releases; only the latest release is supported. Response is best-effort, and an advisory names the affected and fixed versions.

There aren't any published security advisories