Only the latest release on PyPI is supported with fixes.

Please report vulnerabilities privately via GitHub security advisories: Security → Report a vulnerability on this repository. Do not open a public issue for anything exploitable.

RigorLoop executes generated code. Candidate solutions produced by agents (and any custom_python checks you configure) run as subprocesses on your machine with your user's permissions. The built-in mitigations — hard timeouts, output caps, a scratch working directory, no stdin inheritance — are guardrails against runaway code, not a security boundary. If your threat model includes actively malicious generated code, run RigorLoop inside a container or VM. OS-level sandboxing is on the roadmap.

RigorLoop also sends your task description and dev-split example content to the configured model via the claude CLI (validation/test examples are additionally embedded, one at a time, in evaluation calls). Do not put data in your examples file that must not leave your machine.