Review ergonomics
ergonomicsEnsure a home floor plan fits its occupants: check walking space, kitchen layout, wheelchair turning, beds and windows. Receive a habitability score and detailed findings.
Instructions
Ergonomics and habitability review for the people living there. It never blocks anything: it reads the drawing and says what it finds, and the drawing stays the user's — somebody sketching to learn or to see an idea is not stopped by a standard. People (occupants, children, elderly, wheelchair, stature cm, city, scope) given here weigh this review only; set_home(people=…) keeps them with the project, and every review and dry run then scores for them. It covers room to walk beside beds and in front of kitchen equipment, beds/seats/bathrooms/wardrobes per person, kitchen (work triangle, counter heights, Alexander's counter lengths, five work zones, sockets, gas ventilation, extraction), doors, ceiling heights, windows, minimum furniture, wheelchair turning. Reply {score, score_basis, scope, scores:{architecture,electrical,plumbing}, layout, coverage, capacity, findings:[{sev, place, msg, key, weight, discipline, in_scope, src?, fix?, accepted?}], sources:{src:[title, tier, url]}}. scope={electrical:false,plumbing:false} scores architecture only; all findings remain visible, excluded ones weigh zero. layout includes geometric checks on the active storey; coverage declares limits. Scores are heuristic, not project completion or certification. weight is what the score would gain if that one went away, so a score that moved can be read. src is the source the finding stands on, empty when it is common practice; resolve it in sources instead of asking. tier is the reliability ladder A obliges (Brazilian standard, municipal code) · B references (foreign standard) · C doctrine · D measured · E survey — and it is why a finding is an error or only a tip. fix, when present, is a checked change as tool arguments (move or update): apply one, then review again (fixes of one review may overlap). city, e.g. sao-paulo, lets the municipal code judge instead of only advising; against a standard the more restrictive one wins — set it once with set_home(city=…) so dry runs and check_layout weigh the same rules. key names the finding for the accept tool: a finding accepted there stays in the report with its reason and stops costing score, which is what lets a correct plan reach zero pendencies honestly. orphaned [[key, reason]] lists acceptances no current finding answers to — the problem was fixed, and would come back already silenced; accept(prune=true) drops them.
Input Schema
| Name | Required | Description | Default |
|---|---|---|---|
| city | No | City whose building code applies for this call only, e.g. `sao-paulo`; `set_home(city=…)` keeps it with the project | |
| scope | No | Disciplines included in the score. Architecture always counts; excluded findings remain visible | |
| elderly | No | Of them, elderly people | |
| stature | No | Height of the main cook, cm, to size the countertop (default 165) | |
| children | No | Of them, children (sleep in single beds or cribs) | |
| occupants | No | People living in the home (default 2) | |
| wheelchair | No | Someone uses a wheelchair: NBR 9050 turning space, doors and reach |
Output Schema
| Name | Required | Description | Default |
|---|---|---|---|
| scope | No | Disciplines counted: {electrical, plumbing} | |
| score | Yes | Habitability score 0-100 | |
| layout | No | Layout issues of the storey, as check_layout gives them | |
| scores | No | Score by discipline: {architecture, electrical, plumbing} | |
| sources | No | Standards cited, by code: [title, tier, url] | |
| capacity | No | What the home holds: beds, bedrooms, bathrooms, seats, wardrobe cm | |
| coverage | No | What was checked and what was not | |
| findings | Yes | {sev, place, msg, key, weight, discipline, in_scope, src?, fix?, accepted?} | |
| orphaned | No | Acceptances whose finding is gone: [key, reason, successor key?] | |
| score_basis | No | What the score is: habitability_heuristic |