everyone on the crew has one job
point at a figure to see what it does, click it to open its page.tap a figure to read more, tap it again to open its page.
why a crew
-
one job each, done properly
instead of one agent doing everything, each crew agent is set up for a single job, like planning, the backend, the tests or the review, and works through that job the same careful way every time.
-
they hand work to each other
you give your request to mash, the conductor. mash has tally plan it, hands each part to the agent whose job it is, checks what comes back and merges it.
-
the whole team gets the same crew
a team is a few files in your repository, so everyone who pulls has the same agents set up the same way. with agentmash on, each one shows up in the live room as itself.
set it up in a minute
-
pick who plays
open the team builder and start from a preset (feature, bugfix, release, mobile, ai feature) or pick agents one by one. mash is in every team.
-
add them to your repository
run the command the builder gives you in the repository, for example:
npx agentmash team add "my team" --agents tally,pipe,flex,probe,gavelthen commit the files it writes. teammates get the same crew on their next pull.
-
hand it a job
start claude code at the repository root and say
my team, add rate limiting to the login endpoint, or type/team-my-teamand the task.





















