analysis scientist bench

runs the registered analysis

exactly as the plan says, on the data atlas checked, and it reports a failed prediction as plainly as a passed one.

see the whole lab

call it: the plan is approved and locked, or a figure has to regenerate from the results.
it can: read and search the code, run commands, write and edit files, and edit notebooks.model · same as your session
hands off: controls and baselinesfoil

training modelsbatch

the verdictassay

bench comes with the lab how a study runs

claude plugin install agentmash-lab@agentmash

first time on this machine? add the marketplace before it: claude plugin marketplace add agentmash-dev/agentmash-crew

bench at work

what it does, how, what you get back, who it hands off to and when to call it, in 23 seconds.

each one does one job. open one to see when to call it.