Skip to content

Security: stevenzg/tarjan

Security

SECURITY.md

Security Policy

Supported versions

tarjan is pre-1.0 and ships from main. Security fixes are applied to the latest release and the main branch. Older tagged releases do not receive backports — please upgrade to the newest version (tarjan upgrade).

Reporting a vulnerability

Please do not report security vulnerabilities through public GitHub issues, pull requests, or discussions.

Instead, report them privately using GitHub's built-in private vulnerability reporting ("Report a vulnerability" under the repository's Security tab). This creates a private advisory only the maintainers can see.

Please include:

  • A description of the issue and its impact.
  • Steps to reproduce (a minimal tarjan.yaml / tarjan.star and command line where relevant).
  • The tarjan version output and your OS/arch.
  • Any suggested remediation, if you have one.

You can expect an initial acknowledgement within a few days. We will keep you informed of progress toward a fix and coordinate disclosure timing with you.

Scope and threat model

tarjan orchestrates local development environments. Keep the following in mind when assessing impact:

  • It runs code you point it at. tarjan clones repositories and runs the setup/run commands defined in a project's tarjan.yaml or tarjan.star. A malicious config or repository can run arbitrary commands on your machine — the same trust level as running make or npm install in that project. Only run configs you trust.
  • Secrets and env files. tarjan loads env files that are intentionally kept out of git. Reports about secrets being logged, written to world-readable locations, or leaked into child processes unintentionally are in scope.
  • The installer and self-update path. install.sh / install.ps1 and tarjan upgrade download release binaries. Issues that would let an attacker substitute or tamper with a downloaded binary are in scope.
  • Out of scope: vulnerabilities in the third-party services, tools, or repositories a user chooses to orchestrate; social-engineering a user into running an untrusted config.

There aren't any published security advisories