ralph-research is a local-first runtime; there is no hosted service to patch. The current security-supported line is the latest release on main:

If you depend on a pinned older minor, please open an issue so we can decide whether to backport. We will not silently backport without that conversation.

Please do not file public GitHub issues for security problems. Instead:

1. Open a GitHub security advisory at https://github.com/coyaSONG/ralph-research/security/advisories/new.
2. Include a reproduction, the version of ralph-research you tested, your Node.js version, and the operating system.
3. If you cannot use GitHub security advisories, contact the maintainer through the email address on their GitHub profile.

We will acknowledge receipt within 7 days and aim to ship a fix or a documented mitigation within 30 days for issues we agree are exploitable.

• The CLI and MCP server execute commands from the ralph.yaml manifest in the same Git repository as the runtime. Treating a manifest from an untrusted source as safe is out of scope — manifests run arbitrary commands by design. Always read ralph.yaml before running rrx.
• Persisted state under .ralph/ may include LLM judge rationales and experiment outputs. Treat that directory with the same care as any other build artifact your repository produces.
• The runtime never uploads state or telemetry to a remote service.

If you are reporting an issue that also affects upstream dependencies (commander, execa, pino, yaml, zod, @modelcontextprotocol/sdk), please disclose to those projects first or at the same time. We will coordinate release windows where it makes sense.