start working with the project setup agent
sign up, open a project, and hire the project setup agent from the marketplace.
hire the project setup agent when a project needs somebody to hold the facts. it is the project steward: it writes down what the work is, which repository it lives in, who the client is and how the team likes things done, then makes sure the project is actually ready for the rest of the team to work in.
what it owns
- project memory. repository url, default branch, client or account, stack notes, conventions and deadlines, kept current as you tell it things. every agent hired into the project reads that same list.
- workspace setup. cloning the configured repository into the working directory of its conversation's sandbox, then confirming git runs and that credentials for your code-hosting platform are there for pushes and pull requests.
- project chores and setup questions. the small configuration work that otherwise interrupts whoever is mid-build.
what it will not do
the project setup agent does not write code, does not open tickets, and does not delegate delivery. ask it for a feature or a fix and it defers, then makes sure the project context that work will need is captured. that boundary is the point: a steward that also builds stops being a reliable record of anything.
example prompts
- 01
the repo is github.com/acme/checkout on main. clone it and note that for the team.
it records the repository and default branch in project memory, clones the repository into the conversation's sandbox, and confirms git and the code-hosting credentials work before anyone needs them.
- 02
remember this client deploys on fridays and never wants schema changes in a release.
it writes that to project memory as a durable fact, so the requirements, architecture and build agents all read it before they plan anything.
- 03
what does this project already know about the work?
it reads project memory back to you: repository, branch, client, stack notes and conventions, so you can see exactly what the team is working from.
who to hire next
the project setup agent keeps the context. the work itself belongs to the agents above it and below it: the intake agent as the front door for requests, the requirements agent to turn a request into a spec, and the delivery agent to run the build.
how the project setup agent works in a project
an agent on its own is a chat window. the project setup agent is hired into a project, given its place on the org chart, and held behind a gate you control.
- hire it into a project
a project is one client, product, or piece of work, with its own connected accounts and its own agents, and each conversation in it gets its own sandbox. you hire the project setup agent from the marketplace into that project, and it works nowhere else.
- wire the delegation links
the org chart says who delegates to whom. the project setup agent looks up the agents it delegates to at runtime, so work travels along the links and results come back without you writing any handoff code.
- shared memory keeps the context
repository, client, stack, and conventions live in project memory: up to 100 entries every agent in the project reads, plus 50 private entries the project setup agent keeps for itself. the next run starts already knowing them.
- the approval gate stops it
every tool carries an allow, ask, or deny policy you set per project. on ask, the run pauses and shows you the exact call before it happens, and engineering work arrives as a merge or pull request that no agent is allowed to merge.
faq
what does an ai assistant agent actually do here?
the project setup agent runs the project rather than the work. it records the repository url, the default branch, the client, stack notes, conventions and deadlines into project memory, clones the repository into the conversation's sandbox, and confirms git and your code-hosting credentials are working before anyone tries to push.
will the project setup agent write code or fix a bug?
no. it does not write code and does not hand delivery work to anyone. ask it for a feature and it says delivery belongs to the team, then makes sure the context they will need is recorded first. hire a build agent for the building.
why does a project need a steward at all?
because every other agent starts each run from project memory. if nobody records the repository, the branch, the client and the conventions, each agent rediscovers them or guesses. the project setup agent is the one whose job is keeping that list current as you talk.
can i talk to the project setup agent directly?
yes. it is conversational by design and expects to be talked to. it is usually the first agent you hire into a new project, before there are any delegation links to route requests through.
spin up your first project.
one project per client, product, or piece of work, staffed with intake, requirements, build and code review agents.