Tasks · build · property-management
Maintenance app for six apartment buildings with work requests, work orders, vendors, approvals, and a dashboard
property-maintenance · 3600 s per turn · source
The ask
We manage six apartment buildings, and I need one place to run maintenance: tenants' work requests come in, we turn the ones we accept into work orders, and each work order goes to a vendor who does that kind of work and whose insurance is current. Any work order estimated over $1,500 has to wait for my approval before the vendor is sent. Our units, tenants, vendors, and last year's requests are attached; they came out of different systems and repeat themselves in places, so make sure nothing is counted twice. I want team logins with different access: each building manager sees and handles only their own buildings, and our accountant can see everything but change nothing. Give me a dashboard that starts with how many requests are open, an admin login for me, and a login for Rosa Delgado, who manages Harbor View and Elm Court.
Seed files the agent must import: units.csv</code>, <code>tenants.csv</code>, <code>vendors.csv</code>, <code>work_requests.csv. The agent must leave a RESULT.json with a reachable URL, an admin login, and the named restricted login.
Acceptance checklist
26 binary items. The first 14 are the enterprise baseline shared by every build task (delivery, invitations, roles, server-side authorization, scoped live dashboard, exact lists and exports, audit metadata, validation, persistence); the rest are application-specific. Core items, which together define "first usable": 1, 3, 5, 6, 8, 14, 15, 21, 24.
| Tag | Items |
|---|---|
| Exact | 8 |
| Rule | 4 |
| Permission | 3 |
| Automation | 2 |
| Dashboard | 2 |
| Delivery | 2 |
| Sharing | 2 |
| API | 1 |
| Audit | 1 |
| Persistence | 1 |
The full checklist with per-item "how to check" instructions and expected figures is in checklist.md. Testers see it; the agent never does.
Change requests
Each change is a new one-shot turn on the carried-forward workspace. Only the owner's text below is sent; the tester items are withheld. After each change the full checklist is rerun, so regressions count.
Change 1 · 3 new checklist items
I need to see what maintenance is costing us: a report of the cost of completed work per building for each month, and on the dashboard, how many days on average it takes us to finish a request.
Change 2 · 3 new checklist items
When a tenant caused the damage I want them billed: let us mark a work order as tenant-caused, and when that work is completed, put its actual cost on the tenant's balance.
Change 3 · 3 new checklist items
Vendors should log in themselves: each vendor sees only the work orders we've sent them and marks them done with the actual cost. And add a Scheduled step before work starts, with the date of the visit.
No published score. Build runs need a human acceptance pass on a live URL, and no arm has completed one for the full set. Automated probes generate a tester sheet; they do not substitute for it. Results will be published as a separately labeled release when an arm completes.