Workspaces and projects¶
kata separates repository files from issue data. Repositories and workspaces
carry only enough information to resolve a project. The database lives under
KATA_HOME unless KATA_DB points somewhere else.
Initialize¶
In a git workspace, kata init derives the project name from the git remote.
Pass --project to choose the name explicitly:
kata init writes .kata.toml and ensures .kata.local.toml is ignored. The
local file is for per-machine settings such as a remote daemon URL.
Bind many workspaces to one project¶
Use the same project name in each workspace:
cd ~/code/product
kata init --project product
cd ~/code/product-worktree
kata init --project product
Both workspaces now resolve to the same project in the same local kata database. Issue short IDs, labels, links, and events are shared.
Run from outside a workspace¶
Use the global --workspace flag:
This is useful for scripts, cron jobs, and agents that keep their own working directory.
Project commands¶
List and inspect projects:
Rename a project:
Merge accidental duplicates:
Archive and restore a project:
projects remove hides the project from normal resolution but preserves events
for audit.
Detach one alias when a workspace identity was attached to the wrong project:
Use kata projects show <project> before destructive or structural project
operations so you know which project and aliases are affected.
.kata.toml¶
The committed binding file is intentionally small:
Do not put tokens or host-specific daemon URLs in .kata.toml.
.kata.local.toml¶
The local override file is ignored by git. A common use is routing one workspace to a remote daemon:
KATA_SERVER wins over .kata.local.toml when both are set.
Non-git workspaces¶
kata works without git. Use an explicit project name:
The issue model does not depend on git commits. Git is only one way to derive a default project name and one possible source of close evidence.