The default profile for second-brain:
ACE-aligned folders (Atlas/, Efforts/, Cards/, Calendar/, Sources/),
16 Claude Code skills (11 global, 5 vault), zk note templates, and the
brain plugin identity. A brain profile is everything under <brain>/.brain/ —
skills, templates, hooks, and profile.toml.
This is the profile brain-init clones by default — a plain docker run of
the second-brain image gets it with zero configuration (network required on
first init).
To pick a different profile, or a fork of this one:
# Container
docker run -d -e BRAIN_PROFILE_REPO=https://github.com/you/your-profile ... kitchencoder/second-brain:latest
# Host
brain-init --profile-repo https://github.com/you/your-profile /path/to/brain
brain-init --profile-repo /path/to/local/profile-dir /path/to/brain # offline / localFork this repo to make your own flavour: edit profile.toml (folders,
skills, plugin identity, zk conventions) and the skills/, templates/,
hooks/ it references. After seeding, brain-profile update fast-forwards
a brain's clone to this repo's latest. profile.toml declares schema = 1
(the manifest format version); an engine older than a profile's declared
schema refuses to load it rather than misreading it.
The plugin identity (plugin.name, hook marker, MCP server name) is shared by
every brain using this profile. Give each additional brain a name at init —
brain-init --brain-name work (or -e BRAIN_NAME=work) — which qualifies the
staged identity (second-brain-work, brain-work) and remap its published
host ports (--mcp-port / --api-port). See the second-brain
User Guide
for the full two-brain and tunnel/reverse-proxy walkthroughs.