Skip to content

Security: eviweb/ai-rules

Security

SECURITY.md

Security Policy

Supported versions

Only the latest release on main receives security fixes.

Reporting a vulnerability

Do not open a public GitHub issue for security vulnerabilities.

Report vulnerabilities by emailing the maintainer directly:

  • Email: dev@eviweb.fr
  • Subject: [ai-rules] Security vulnerability

Please include:

  • A description of the vulnerability and its potential impact
  • Steps to reproduce or a proof-of-concept
  • Any suggested remediation if you have one

You will receive an acknowledgement within 48 hours and a status update within 7 days.

Scope

This project is a local CLI tool that manages symlinks and configuration files on the user's own machine. It does not handle credentials, network requests, or multi-user environments. The primary risk surface is:

  • Symlink injection via a malicious agents.toml
  • TOML config patching writing unexpected values to user config files

Out of scope

  • Vulnerabilities in third-party AI assistants (Claude Code, Codex, Gemini)
  • Issues that require physical access to the user's machine

There aren’t any published security advisories