Skip to content

Security: tlbx-ai/tlbx

SECURITY.md

Security policy

Report a vulnerability privately

Use GitHub's private vulnerability reporting form:

Report a vulnerability

Do not open a public issue for an undisclosed vulnerability. Include:

  • the affected tlbx version, update channel, and platform;
  • the security impact and conditions required to reproduce it;
  • reproduction steps or a minimal proof of concept;
  • any known exploitation or suggested mitigation.

Remove passwords, API keys, tokens, personal data, and unrelated repository content from the report. For security hardening ideas that do not disclose an exploitable vulnerability, use the normal public issue tracker.

What to expect

The project aims to acknowledge a complete report within three business days and provide an initial assessment within seven business days. Complex reports may take longer to reproduce or remediate. The reporter and maintainer should coordinate publication so users have a practical mitigation or update before technical details become public.

When a vulnerability affects users, tlbx publishes the appropriate combination of a GitHub Security Advisory, release notes, mitigation guidance, and a fixed release. Security updates are provided without an additional fee.

Supported versions

The current stable release is the supported production line. Reports against a dev prerelease are accepted, but prereleases are evaluation builds and may require updating to a newer prerelease rather than receiving a backport. Historical releases do not receive separate fixes unless a security advisory explicitly says otherwise.

See SUPPORT.md for the lifecycle policy.

Release integrity

Native release archives are accompanied by a platform-specific SPDX SBOM and GitHub build-provenance and SBOM attestations. Installers and the built-in updater additionally require a signed manifest that binds the release version, platform, channel, and packaged-file hashes.

See Release integrity for verification steps.

Security contact discovery

The canonical human-readable policy is also published at tlbx.ai/security. Automated security-contact discovery is available at tlbx.ai/.well-known/security.txt.

There aren't any published security advisories