RWM is a developer tool, not a sandbox. It assumes the workspace it operates on is code you already trust to run on your machine.

Specifically:

• Gemfile parsing. RWM reads each package's Gemfile via Bundler's DSL (Bundler::Dsl#eval_gemfile), which evaluates Gemfile contents as Ruby. Parsing a malicious Gemfile is arbitrary code execution.
• Rake tasks. rwm run, rwm spec, and rwm bootstrap shell out to rake and bundle in each package directory. Running rake against a malicious Rakefile is arbitrary code execution.
• Path resolution. rwm resolves the workspace root via git rev-parse --show-toplevel and reads files beneath it. Symlinks are followed.

This is the same trust posture as bundle install or rake itself — if you'd already run those in a directory, you can run rwm there.

Do not run rwm in directories containing untrusted Gemfiles or Rakefiles (for example, a fresh clone of an unfamiliar repository you haven't reviewed).

Until we reach 1.0, only the latest release receives security patches. There are no backports to older versions.

If you want security fixes as soon as they ship, leave the version unconstrained:

gem "ruby_workspace_manager"

If you prefer stability and can tolerate a short delay before upgrading, pin to the current minor:

gem "ruby_workspace_manager", "~> 0.6"

Everything can break between minor versions before 1.0. This policy will be revisited after 1.0.

If you discover a security vulnerability, please report it privately via GitHub Security Advisories.

Do not open a public issue for security vulnerabilities.

You can expect an initial response within 72 hours. Once confirmed, a fix will be prioritized and released as a patch version.

Releases are pushed to RubyGems manually by the maintainer with MFA enabled. This is intentional — there is no CI/CD path to RubyGems, so a compromised pipeline cannot publish a malicious release.