Skip to content

Security: rennzhang/ai-dispatch

Security

SECURITY.md

Security

Do not report secrets in public issues.

ai-dispatch shells out to local AI coding CLIs. Those CLIs may read files, edit files, or run commands depending on their own configuration and the flags used by ai-dispatch.

Before using ai-dispatch in automation:

  • use a workspace intended for agent execution
  • avoid mounting secrets that the provider does not need
  • keep notification hooks free of prompt text, response text, secrets, full stderr, and private paths
  • review provider CLI authentication and permission settings

Report sensitive issues privately to the repository owner.

There aren't any published security advisories