Skip to content

Security: brolag/neural-codex

Security

docs/SECURITY.md

Security

Guardrails

  • Keep network allowlists narrow.
  • Treat external tool output as untrusted.
  • Never commit secrets.

Sensitive Changes

  • Require explicit approval for auth, data access, or infra changes.
  • Document threat model in the relevant design doc.

There aren’t any published security advisories