Skip to content

Security: mikespax/codex-usage-remote

Security

SECURITY.md

Security model

Codex Usage Remote is deliberately read-only.

  • It calls only account/rateLimits/read through codex app-server.
  • It never reads or stores ~/.codex/auth.json, OAuth tokens, API keys, browser cookies, ChatGPT web sessions, prompts, or conversation/session data.
  • It never uploads usage data and has no developer-controlled backend or analytics.
  • SSH private keys are never opened, copied, or managed. /usr/bin/ssh uses the user's existing agent and ~/.ssh/config.
  • SSH host and executable values are validated and passed as separate process arguments. Shell fragments, whitespace, redirects, and command substitution are rejected.
  • Operational errors are shown without displaying subprocess stderr, which can contain sensitive host paths or remote output.
  • Preferences are local UserDefaults; rate-limit results are kept in memory and are not persisted.

Report suspected vulnerabilities privately to the repository owner before public disclosure.

There aren't any published security advisories