Skip to content

Security: omkarchandra/ocdeck

SECURITY.md

Security policy

OC Deck handles agent permissions and launches agent processes, so please report security problems privately.

Reporting a vulnerability

Use GitHub's private vulnerability reporting (Security tab → Report a vulnerability) on this repository. If that is unavailable, open a minimal public issue that says only that you have a security report and ask for a private channel. Please do not include exploit details in public.

Helpful details: the version, the harness involved, and the steps to reproduce. I aim to acknowledge reports within a week.

What is in scope

  • OC Deck approving, denying or starting something the user did not ask for.
  • Secrets (tokens, keys, passwords) reaching a child process, a log, a state file or the terminal.
  • Reading or writing outside the files OC Deck is documented to use.
  • The optional GNOME extensions, plugins and agent browser in this repository.

Design notes

OC Deck never approves a permission by itself; every approval is a keystroke from the user. It removes server credentials from the environment of processes it starts, bounds and validates each file it reads, and refuses to replace launchers it did not create. The agent browser's grant file records a choice but is not a security boundary; enforce browser access in the browser tool's own configuration.

There aren't any published security advisories