setup-customize

Name: setup-customize
Author: dcb

111stars22forks

How to install setup-customize?

dcb/homeassistant-claude-kit/setup-customize

$npx -y skills add dcb/homeassistant-claude-kit --skill setup-customize

Installs into the current project.

Summary

Run after setup-infrastructure to map rooms, entities, and preferences to the dashboard and automation templates. Conversational and resumable. Trigger phrases: "customize my home", "set up rooms", "configure automations", "map my entities", "set up my dashboard", "finish setup",

Files · 1

View on GitHub

SKILL.md

1	# Setup Customize
2
3	This skill maps your Home Assistant instance to the dashboard and automation templates
4	through a guided interview. It is resumable — if the conversation ends mid-way,
5	re-invoke this skill and it will pick up from the last checkpoint.
6
7	See `references/question-patterns.md` for detailed question wording and example answers
8	for each domain.
9
10	## Step 0: Check Prerequisites
11
12	Verify `setup-state.json` exists and infrastructure is complete:
13
14	```python
15	import json, sys, os
16	if not os.path.exists('setup-state.json'):
17	print('NOT_READY'); sys.exit(0)
18	with open('setup-state.json') as f:
19	state = json.load(f)
20	schema = state.get('schema_version', 0)
21	if schema > 1:
22	print('SCHEMA_WARNING')
23	phase = state.get('session', {}).get('current_phase', '')
24	infra = state.get('infrastructure', {}).get('steps_completed', [])
25	if 'infrastructure_complete' in phase or 'pull' in infra:
26	answers = state.get('answers', {})
27	if answers.get('rooms') or phase.startswith('customize:'):
28	print('RESUME')
29	print(f'PHASE:{phase}')
30	print(f'ROOMS_DONE:{",".join(answers.get("rooms", {}).keys())}')
31	else:
32	print('FRESH')
33	else:
34	print('NOT_READY')
35	```
36
37	Run via `python3 -c "..."` and check the output:
38
39	- `NOT_READY`: Tell user to run `setup-infrastructure` first.
40	- `SCHEMA_WARNING`: State file from newer version — proceed with caution.
41	- `RESUME`: Load checkpoint. Tell user: "Welcome back! You were at [phase]. Rooms done: [list]. Continuing."
42	- `FRESH`: Begin from Phase 1.
43
44	### Checkpoint Writing Pattern
45
46	After EVERY user answer, update `setup-state.json` with granular progress:
47
48	```python
49	import json
50	def save_checkpoint(phase, answers_update=None, files_written=None):
51	with open('setup-state.json') as f:
52	state = json.load(f)
53	state['session']['current_phase'] = phase
54	if answers_update:
55	state.setdefault('answers', {}).update(answers_update)
56	if files_written:
57	state.setdefault('files_written', []).extend(files_written)
58	with open('setup-state.json', 'w') as f:
59	json.dump(state, f, indent=2)
60	```
61
62	Example calls:
63	- `save_checkpoint('customize:room_mapping', {'rooms': {'living_room': {'light': '...', 'motion': '...'}}})`
64	- `save_checkpoint('customize:domain_selection', {'domains_selected': ['lighting', 'climate']})`
65	- `save_checkpoint('customize:notifications', {'notify_targets': {'primary': 'notify.mobile_app_x'}})`
66	- `save_checkpoint('customize:files', files_written=['config/automations/lighting.yaml'])`
67
68	## Step 1: Discover Entity + Area + Floor Registries
69
70	Primary method: Use registry data. Entity-to-room assignment should come from the
71	device/entity registries (via `area_id`) whenever possible. This is the authoritative source.
72
73	Fallback: Name inference + user confirmation. If the registries have sparse area
74	assignments (common in setups where the user hasn't organized areas in HA), you may infer
75	room assignments from entity ID naming patterns (e.g., `bedroom_motion` → bedroom).
76	However, when using name inference, you MUST:
77	1. Clearly mark inferred assignments as "inferred (not in registry)"
78	2. Ask the user to confirm ALL inferred assignments before proceeding
79	3. Never present inferred data as verified fact
80
81	### 1a. Query Floor + Area Registries
82
83	Get the authoritative room and floor structure:
84
85	```bash
86	source .env && ssh "$SSH_USER@$HA_HOST" "source /etc/profile.d/claude-ha.sh; source ${HA_REMOTE_PATH:=/config/}.env; ha-ws raw config/floor_registry/list" 2>/dev/null
87	source .env && ssh "$SSH_USER@$HA_HOST" "source /etc/profile.d/claude-ha.sh; source ${HA_REMOTE_PATH}.env; ha-ws raw config/area_registry/list" 2>/dev/null
88	```
89
90	This gives you:
91	- All floors with IDs and names
92	- All areas with `floor_id` assignments
93	- Do NOT ask the user about floors if this data is available.
94
95	### 1b. Query Device + Entity Registries
96
97	Get the authoritative entity-to-area mappings:
98
99	```bash
100	source .env && ssh "$SSH_USER@$HA_HOST" "source /etc/profile.d/claude-ha.sh; source ${HA_REMOTE_PATH}.env; ha-ws raw config/device_registry/list" 2>/dev/null
101	source .env && ssh "$SSH_USER@$HA_HOST" "source /etc/profile.d/claude-ha.sh; source ${HA_REMOTE_PATH}.env; ha-ws raw config/entity_registry/list" 2>/dev/null
102	```
103
104	Entity-to-area resolution chain:
105	1. Check `entity_registry` → if the entity has a direct `area_id`, use it
106	2. Otherwise, find the entity's `device_id` → look up that device in `device_registry` → use the device's `area_id`
107	3. If neither has an `area_id`, the entity is unassigned — note it but do NOT guess
108
109	### 1c. Query Entities by Domain
110
111	For each relevant domain, query the live en

People who liked this also viewed

6 picks

Type

setup-customize

111stars22forks

View on GitHub

How to install setup-customize?

dcb/homeassistant-claude-kit/setup-customize

$npx -y skills add dcb/homeassistant-claude-kit --skill setup-customize

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# Setup Customize
2
3	This skill maps your Home Assistant instance to the dashboard and automation templates
4	through a guided interview. It is resumable — if the conversation ends mid-way,
5	re-invoke this skill and it will pick up from the last checkpoint.
6
7	See `references/question-patterns.md` for detailed question wording and example answers
8	for each domain.
9
10	## Step 0: Check Prerequisites
11
12	Verify `setup-state.json` exists and infrastructure is complete:
13
14	```python
15	import json, sys, os
16	if not os.path.exists('setup-state.json'):
17	print('NOT_READY'); sys.exit(0)
18	with open('setup-state.json') as f:
19	state = json.load(f)
20	schema = state.get('schema_version', 0)
21	if schema > 1:
22	print('SCHEMA_WARNING')
23	phase = state.get('session', {}).get('current_phase', '')
24	infra = state.get('infrastructure', {}).get('steps_completed', [])
25	if 'infrastructure_complete' in phase or 'pull' in infra:
26	answers = state.get('answers', {})
27	if answers.get('rooms') or phase.startswith('customize:'):
28	print('RESUME')
29	print(f'PHASE:{phase}')
30	print(f'ROOMS_DONE:{",".join(answers.get("rooms", {}).keys())}')
31	else:
32	print('FRESH')
33	else:
34	print('NOT_READY')
35	```
36
37	Run via `python3 -c "..."` and check the output:
38
39	- `NOT_READY`: Tell user to run `setup-infrastructure` first.
40	- `SCHEMA_WARNING`: State file from newer version — proceed with caution.
41	- `RESUME`: Load checkpoint. Tell user: "Welcome back! You were at [phase]. Rooms done: [list]. Continuing."
42	- `FRESH`: Begin from Phase 1.
43
44	### Checkpoint Writing Pattern
45
46	After EVERY user answer, update `setup-state.json` with granular progress:
47
48	```python
49	import json
50	def save_checkpoint(phase, answers_update=None, files_written=None):
51	with open('setup-state.json') as f:
52	state = json.load(f)
53	state['session']['current_phase'] = phase
54	if answers_update:
55	state.setdefault('answers', {}).update(answers_update)
56	if files_written:
57	state.setdefault('files_written', []).extend(files_written)
58	with open('setup-state.json', 'w') as f:
59	json.dump(state, f, indent=2)
60	```
61
62	Example calls:
63	- `save_checkpoint('customize:room_mapping', {'rooms': {'living_room': {'light': '...', 'motion': '...'}}})`
64	- `save_checkpoint('customize:domain_selection', {'domains_selected': ['lighting', 'climate']})`
65	- `save_checkpoint('customize:notifications', {'notify_targets': {'primary': 'notify.mobile_app_x'}})`
66	- `save_checkpoint('customize:files', files_written=['config/automations/lighting.yaml'])`
67
68	## Step 1: Discover Entity + Area + Floor Registries
69
70	Primary method: Use registry data. Entity-to-room assignment should come from the
71	device/entity registries (via `area_id`) whenever possible. This is the authoritative source.
72
73	Fallback: Name inference + user confirmation. If the registries have sparse area
74	assignments (common in setups where the user hasn't organized areas in HA), you may infer
75	room assignments from entity ID naming patterns (e.g., `bedroom_motion` → bedroom).
76	However, when using name inference, you MUST:
77	1. Clearly mark inferred assignments as "inferred (not in registry)"
78	2. Ask the user to confirm ALL inferred assignments before proceeding
79	3. Never present inferred data as verified fact
80
81	### 1a. Query Floor + Area Registries
82
83	Get the authoritative room and floor structure:
84
85	```bash
86	source .env && ssh "$SSH_USER@$HA_HOST" "source /etc/profile.d/claude-ha.sh; source ${HA_REMOTE_PATH:=/config/}.env; ha-ws raw config/floor_registry/list" 2>/dev/null
87	source .env && ssh "$SSH_USER@$HA_HOST" "source /etc/profile.d/claude-ha.sh; source ${HA_REMOTE_PATH}.env; ha-ws raw config/area_registry/list" 2>/dev/null
88	```
89
90	This gives you:
91	- All floors with IDs and names
92	- All areas with `floor_id` assignments
93	- Do NOT ask the user about floors if this data is available.
94
95	### 1b. Query Device + Entity Registries
96
97	Get the authoritative entity-to-area mappings:
98
99	```bash
100	source .env && ssh "$SSH_USER@$HA_HOST" "source /etc/profile.d/claude-ha.sh; source ${HA_REMOTE_PATH}.env; ha-ws raw config/device_registry/list" 2>/dev/null
101	source .env && ssh "$SSH_USER@$HA_HOST" "source /etc/profile.d/claude-ha.sh; source ${HA_REMOTE_PATH}.env; ha-ws raw config/entity_registry/list" 2>/dev/null
102	```
103
104	Entity-to-area resolution chain:
105	1. Check `entity_registry` → if the entity has a direct `area_id`, use it
106	2. Otherwise, find the entity's `device_id` → look up that device in `device_registry` → use the device's `area_id`
107	3. If neither has an `area_id`, the entity is unassigned — note it but do NOT guess
108
109	### 1c. Query Entities by Domain
110
111	For each relevant domain, query the live en

People who liked this also viewed

6 picks

Type