Skip to main content

corgi agent

corgi agent

Keep Claude Code Remote Control running for your corgi workspaces

Synopsis

Agent mode makes this machine an always-on, multi-repo Remote Control host.

Remote Control already gives you a phone-driven Claude Code session on your own machine. Two things stop it being always-on: the local process must keep running, and it exits after roughly ten minutes awake without network.

corgi agent supervises it — restarting after a network timeout, holding a wake lock so the machine does not sleep mid-session, and running one per workspace under that workspace's own Claude config directory.

Getting started: corgi agent init # in a stack, once corgi agent install # start at login corgi agent status # what is running, and under which account

Options

-h, --help help for agent

Options inherited from parent commands

--describe Describe contents of corgi-compose file
--dockerContext string Specify docker context to use, can be default,orbctl,colima (default "default")
-l, --exampleList List examples to choose from. Click on any example to download it
-f, --filename string Custom filepath for for corgi-compose
--fromScratch Clean corgi_services folder before running
-t, --fromTemplate string Create corgi service from template url
--fromTemplateName string Create corgi service from template name and url
-g, --global Use global path to one of the services
--interactive Force interactive prompts even when no TTY/agent detected
--isolate string Run this workspace under a named lease: own port block, own database names, own containers
--json Emit machine-readable JSON output
--privateToken string Private token for private repositories to download files
-o, --runOnce Run corgi once and exit
--silent Hide all welcome messages

SEE ALSO

Auto generated by spf13/cobra on 31-Aug-2026