start working with the intake agent
sign up, open a project, and hire the intake agent from the marketplace.
hire the intake agent when you want one place to send requests. it is the front door of the project: you describe what you want in plain language, it works out who should handle it, and it comes back to you with the result rather than with a status meeting.
what it owns
- intake. your request, taken as stated, without you having to know which agent should get it.
- routing. requirements go to whichever delegate handles business analysis; delivery goes to whichever delegate runs engineering, with the spec path passed along. it resolves those delegates at runtime from its own delegation links, so the routing follows your org chart rather than a hardcoded name.
- reporting back. when engineering reports delivery done, it relays that to you with the merge or pull request link, writes release notes into the workspace, and records a short release summary in project memory.
what it will not do
the intake agent never writes code, never creates or scaffolds files, and never commits. it also stopped authoring the requirements doc: that belongs to the requirements agent, which asks you the questions first. if the project has no requirements agent or no delivery agent hired, it says so instead of doing that job badly itself.
example prompts
- 01
we need saved payment methods on checkout. take it from here.
it routes the request to the requirements agent for clarifying questions and a spec, waits, then hands the finished spec to the delivery agent to drive the build.
- 02
what is the status of the checkout work?
it answers from its own record of what it routed and what came back, including the spec path and the pull request link once delivery reports in.
- 03
the pull request is up. write the release notes.
it writes release notes to a docs file in the workspace and records a short release summary in project memory for the next request to build on.
who to hire with it
the intake agent is a router, so it needs somewhere to route to. pair it with the requirements agent for the spec and the delivery agent to run the build, and let those two delegate to their own team.
how the intake agent works in a project
an agent on its own is a chat window. the intake 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 intake 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 intake 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 intake 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 project manager actually do here?
the intake agent is the front door. you give it a feature or a fix in plain language, it hands requirements to whichever delegate does business analysis, hands the finished spec to whichever delegate runs engineering, and relays the result back with the merge or pull request link attached.
does the intake agent write the spec itself?
no. it routes requirements to a business-analysis delegate, which runs the clarifying questions with you and writes the spec. if no requirements agent is hired in the project, it tells you that rather than authoring the spec itself or guessing at scope.
does the intake agent write code?
never. it does not write code, create files, or commit. it routes work by capability and reports results. when it hands delivery over it asks that agent to drive its own team, not to personally write the feature.
what do i get back at the end?
a summary in the chat with the merge or pull request link, release notes written to a docs file in the workspace, and a short release summary recorded in project memory so the next request starts from what already shipped.
spin up your first project.
one project per client, product, or piece of work, staffed with intake, requirements, build and code review agents.