Skip to content

Security: Astralchemist/rig

Security

SECURITY.md

Security Policy

Rig reads your entire codebase to build its graph and runs a local server that exposes it over MCP and HTTP. We take that trust seriously.

What Rig does with your data

  • Everything is local. The graph lives in .rig/rig.db inside your project. Embeddings run on-device. No code, query, or graph data is uploaded anywhere.
  • The web map binds to loopback (127.0.0.1) by default. Binding to a non-loopback address requires a bearer token (--auth); Rig refuses to bind externally without one.
  • rig install writes MCP server entries into your agent's config files (e.g. .claude.json). It merges, never clobbers.
  • The first run downloads an embedding model from the Hugging Face CDN to your local model cache. This is the only outbound network call in the core flow.

Supported versions

Until a 1.0 release, only the latest published version receives security fixes.

Reporting a vulnerability

Please do not open a public issue for security reports.

Email kingolajohn55@gmail.com with:

  • a description of the issue and its impact,
  • steps to reproduce (a minimal repro is ideal),
  • any suggested remediation.

We aim to acknowledge within 72 hours and to ship a fix or mitigation before public disclosure. We'll credit you in the changelog unless you'd prefer to remain anonymous.

Scope

In scope: the rig CLI, the MCP server, the web server/auth, the installer's config writes, and anything that could exfiltrate local code or graph data.

Out of scope: vulnerabilities in upstream dependencies (report those upstream), and issues requiring a pre-compromised local machine.

There aren't any published security advisories