Every day

talea sync                   # clone the new, fast-forward the rest
talea status                 # branch, clean/dirty, ahead/behind, in one table
cd $(talea where eklavya)

`talea sync`

Clones what is missing, fast-forwards what is there. Safe to run unattended, and safe to run twice.

What it will not do, and this is the whole of it:

It will not merge a divergence. Fast-forwards only. Anything else is reported for you to deal with.
It will not move you off your branch. It fast-forwards the branch you are on, or leaves it alone and says why. A branch with no upstream is fetched and left.
It will not touch a dirty repo. Uncommitted changes mean it fetches and stops, with a line saying so.
It will not push.

Every command runs from anywhere. Inside a workspace, that workspace. Outside one, the workspace talea init made — and if this machine has several, it asks which. With no terminal to ask on (a script, a pipe) it lists them and exits non-zero rather than picking one for you.

talea sync -g nonstopio      # one group
talea sync eklavya           # one repo — same as -r eklavya
talea sync --no-clone        # fast-forward only, clone nothing
talea sync --no-adopt        # skip the existing-checkout scan

talea clone is the first half on its own — clone what is missing and never fetch or merge. It is what you want on a slow connection.

`talea status`

Column	What it shows
`REPO`	repository name
`BRANCH`	the branch checked out right now
`DEFAULT`	the repo’s default branch, blank when you are already on it
`STATE`	clean or dirty, and commits ahead or behind origin

talea status --drift         # only repos on some other branch
talea status --missing       # only repos not cloned yet
talea status --all           # the whole catalogue, not just this machine

`talea where`

talea where eklavya          # the absolute path, one line, nothing else
cd $(talea where eklavya)    # what it is actually for
talea where                  # the workspace root

The path goes to stdout on its own, so it composes with cd, code, open and anything else that takes a directory. Every diagnostic goes to stderr, and an unknown name exits non-zero — so cd $(talea where typo) fails rather than landing you in your home directory.

If two owners have a repo of the same name it says so and exits non-zero rather than picking one. Pass owner/name.

Changing what this machine keeps

talea add some-repo          # keep one more here, and clone it now
talea rm some-repo           # take it off the list
talea select                 # reopen the whole checklist
talea pick some-repo         # keep that one — an unknown name opens the checklist

talea rm removes the repo from this machine’s list and leaves the checkout exactly where it is, printing the path. Deleting it is your call — rm -rf already exists for when you mean it.

talea select opens the checklist with your current selection ticked. Everything in the catalogue is listed, including archived repos and forks: this machine is allowed to keep something the default set does not have. --no-clone saves the list without filling it.

talea pick <name> is talea add when the name matches exactly one repo — case does not matter, and owner/name works. When it matches none, or two owners share it, the checklist opens instead, so a typo lands you in the list rather than on an error.

Running one command everywhere

talea exec -- git log --oneline -1
talea exec -g nonstopio -- npm install
talea exec --all -- git branch --show-current

Everything after -- is the command, run once per repo with that repo as the working directory. Repos are processed one at a time so the output stays readable and attributable.

Exit codes

Every bulk command exits non-zero if any repo failed. Per-repo errors are printed and the run keeps going, so without checking the exit code a partial failure reads as success to a script. A skip is not a failure — a sync over a workspace you are working in still exits 0.

Next: Adopting what you have.