Universal setup for Claude Code and OpenAI Codex CLI — security hooks, notifications, status line, and skills.
$ npx -y skills add DefaultPerson/agent-setup --agent claude-code
What's inside
Universal setup for Claude Code and OpenAI Codex CLI — security hooks, notifications, status line, and skills.
rm -rf ~, git push --force, private key reads, archiving ~/.ssh) with a reason instead of a prompt, so unattended runs don't stall; also catches a pkill -f that would kill the agent's own shell, and systemctl stop|restart of units listed in GUARD_PROTECTED_UNITScommit («сделай коммит и пуш» chains into a PR), push-and-pr, /publish, /repo-context; for research use Claude Code's built-in /deep-research[!TIP] Disable desktop notifications — set
DESKTOP_NOTIFICATIONS=0in~/.claude/settings.json(envsection); for Codex, export it in the shell that starts Codex (hooks get Codex's own environment, not[shell_environment_policy]). Audio TTS keeps working.
[!TIP] Remote monitoring and control — use
herdrwith herdrgram to bridge your terminal multiplexer to Telegram, so you can watch and steer your AI coding agents from your phone while keeping full terminal access.
MIT
FAQ
agent-setup is a Claude Code plugin with 4 hand-picked skills for development work, indexed on Flowy. Install it with the command on its page. It includes commit, publish, push-and-pr. Its skills do not fire on their own yet. Request auto-invocation to have Flowy route them as you prompt. Free and open source.
Is this plugin yours?
Claim it with GitHubSubmit a pluginPromote it