| name | plurio-intake-configurator |
| description | Configure @plurio/intake for any website. Fetches the latest configuration reference from GitHub, conducts a full interview covering all parameters, and produces either a standalone HTML snippet or a complete GTM container JSON. Triggers on: "configure Intake", "install Intake", "add Intake to site", "set up @plurio/intake", "Intake snippet", "Intake GTM container", "настроить Intake", "подключить Intake", "скрипт для сайта Intake", "добавить Intake на сайт". |
Plurio Intake Configurator
Configure @plurio/intake for any website through a full interview covering all available parameters, then generate a ready-to-use snippet or GTM container JSON.
| Deliverable | Format | How to deploy |
|---|
| Standalone snippet | intake-snippet.html | Paste the two <script> tags into every page's <head> before </head> |
| GTM container | intake-gtm-container.json | GTM Admin → Import Container → Merge (any workspace) |
Step 0 — Locate script + fetch live config reference
Resolve the path to generate.py — it lives next to this SKILL.md in tools/configurator/:
SCRIPT_DIR="$(cd "$(dirname "$0")" && pwd)"
SCRIPT_DIR="tools/configurator"
Then use WebFetch to retrieve the live configuration reference:
URL: https://raw.githubusercontent.com/plurio/Intake/main/CONFIGURATION.md
If fetch succeeds: use the fetched parameter list as the authoritative reference. If it contains parameters not listed in this skill's interview below, add questions for those parameters before Step 3.
If fetch fails: proceed using the full parameter list defined in this skill (below). No configuration will be missed — all known parameters are covered.
Step 1 — Mode selection
What do you need to generate?
- A — Standalone snippet — two
<script> tags to paste directly into every page's <head>
- B — GTM container — a JSON file to import into Google Tag Manager (Admin → Import Container → Merge)
Save answer to mode (standalone | gtm).
Step 1.5 — Quick or full setup?
How would you like to configure Intake?
- Quick setup ← recommended — 4 essential questions, done in under a minute. Sensible defaults are applied for everything else.
- Full setup — go through every available parameter across 8 groups for complete control.
- Quick setup → go to Step 2-Quick, then skip to Step 3
- Full setup → go to Step 2-Full, then proceed to Step 3
Step 2-Quick — Essential questions only (4 questions)
Ask one at a time. Save answers to /tmp/intake_configurator.json.
Q1 — Consent mode (GDPR)
Does the site use a CMP (OneTrust, Cookiebot, Didomi, Axeptio, or similar) to manage user consent?
- Yes →
"consent_mode": true, "consent_default": "denied", "url_passthrough": true
- Ask: Denied or granted as default? (default:
denied)
- Ask: Pass UTM params through when consent is declined? Recommended: Yes →
"url_passthrough": true/false
- No →
"consent_mode": false
Q2 — Email and phone collection
Should Intake collect and hash email addresses and phone numbers for CRM matching?
Recommended: Yes.
- Yes →
"pii_collection_enabled": true
- No →
"pii_collection_enabled": false
Q3 — User ID
Authenticated users with a user ID? If yes — source: dataLayer / cookie / localStorage?
- No →
"user_id_source": null
- Yes → set
"user_id_source" + "user_id_key"
Q4 — SPA
Single-page app (React, Vue, Angular, Next.js with client-side routing)?
- Yes →
"spa": true
- No →
"spa": false
Defaults applied silently: data_layer: true, analytics_ids: { google_analytics: true }, 7 social referrals (added by the configurator — not a library default), in-app browser detection enabled.
→ Proceed to Step 3.
Step 2-Full — Full interview
Save all answers to /tmp/intake_configurator.json. Ask questions one group at a time — introduce each group with its heading, then ask all questions in that group before moving to the next.
Group A — Privacy & Consent
A1 — Consent mode (GDPR)
Does the site use a CMP (OneTrust, Cookiebot, Didomi, Axeptio, or similar) to manage user consent?
- Yes →
"consent_mode": true; ask A1a and A1b
- No →
"consent_mode": false
A1a — Default consent status
What should the default consent status be before the user makes a choice?
denied — recommended for GDPR sitesgranted — for non-GDPR sites
Save: "consent_default" ("denied" | "granted", default: "denied")
A1b — URL pass-through
Should UTM parameters be passed through for attribution when a user declines consent?
Recommended: Yes — preserves traffic source data even when consent is declined.
- Yes (recommended) →
"url_passthrough": true
- No →
"url_passthrough": false
A1c — Custom consent event names
Does your CMP fire custom GTM events for consent updates (e.g. "OneTrustGroupsUpdated")? If yes, list them. If no — skip.
- If provided:
"consent_event_names": ["<event1>", ...]
- If skipped: omit
A2 — Email and phone collection
Should Intake collect email addresses and phone numbers for CRM matching?
Values are hashed via SHA-256 — raw data is never stored.
Recommended: Yes.
- Yes (recommended) →
"pii_collection_enabled": true
- No →
"pii_collection_enabled": false
A2a — Custom email selectors (only if A2 = Yes)
Intake auto-detects standard email inputs. Do you have custom selectors to add (e.g. "input[name='email_address']")? Skip to use defaults.
- If provided:
"pii_email_selectors": ["<selector>", ...]
- If skipped: omit
A2b — Custom phone selectors (only if A2 = Yes)
Same for phone inputs — any custom selectors? Skip to use defaults.
- If provided:
"pii_phone_selectors": ["<selector>", ...]
- If skipped: omit
Group B — User Identification
B1 — User ID
Does the site have authenticated users with a user ID (CRM, backend, auth system)?
If yes — where should Intake read the ID from?
- No / no authentication →
"user_id_source": null
- Yes, from dataLayer → ask key name (e.g.
userId) → "user_id_source": "dataLayer", "user_id_key": "<key>"
- Yes, from cookie → ask cookie name →
"user_id_source": "cookie", "user_id_key": "<name>"
- Yes, from localStorage → ask key name →
"user_id_source": "localStorage", "user_id_key": "<key>"
B2 — User IP
Should Intake receive the visitor's IP address (passed from your backend)?
Leave blank if not applicable.
- If provided:
"user_ip": "<value or JS expression>"
- If skipped: omit
B3 — Promocodes
Should Intake generate random promo codes for visitors?
- No (default) → omit
- Yes, simple →
"promocode": true
- Yes, with custom settings → ask for config object (length, charset, prefix, etc.) →
"promocode": { ... }
Group C — SPA & GTM Integration
C1 — Single Page App
Is this a single-page app (React, Vue, Angular, Next.js with client-side routing)?
- Yes →
"spa": true
- No →
"spa": false
C2 — GTM dataLayer push
Should Intake push an intk_ready event to GTM's dataLayer after initialization?
Recommended: Yes (required if using GTM triggers based on Intake data).
- Yes (recommended, default) →
"data_layer": true
- No →
"data_layer": false
C3 — GA4 Client ID & Session ID
Should Intake collect GA4 Client ID and Session ID for BigQuery stitching?
Recommended: Yes.
- Yes (recommended, default) →
"analytics_ga4": true
- No →
"analytics_ga4": false
Group D — Traffic Source Classification
D1 — Custom referral sources
This configurator adds 7 social networks (Facebook, Instagram, LinkedIn, Twitter/X, YouTube, TikTok) so their traffic is classified as medium: social (the library's own defaults are only t.co and plus.url.google.com). Do you need to add more referral sources?
- No → use defaults only
- Yes → for each additional source, ask: host, medium, display name
Save as:
"custom_referrals": [{ "host": "...", "medium": "...", "display": "..." }, ...]
D2 — Custom organic (search) sources
Intake has a built-in list of search engines. Do you need to add custom organic sources?
- No → omit
organics
- Yes → for each: host, query param name, display name
Save as:
"organics": [{ "host": "...", "param": "...", "display": "..." }, ...]
D3 — Direct traffic label
What label should Intake use for direct traffic (no referrer, no UTM)?
Default: source: "(direct)", medium: "(none)". Skip to keep defaults.
- If customized:
"typein_attributes": { "source": "...", "medium": "..." }
- If skipped: omit
D4 — Custom UTM fallback parameters
Does the site use non-standard URL parameters as fallbacks for UTM values?
For example, a ref parameter that maps to campaign. Skip if not applicable.
- campaign fallback →
"campaign_param": "<param_name>"
- term fallback →
"term_param": "<param_name>"
- content fallback →
"content_param": "<param_name>"
- If all skipped: omit all three
Group E — In-App Browser Detection
E1 — In-app browser detection
Should Intake detect in-app browsers (Facebook, Instagram, TikTok webviews) and tag them separately?
Recommended: Yes — prevents inflated direct/referral traffic from social apps.
- Yes (recommended, default) → omit (enabled by default)
- No →
"in_app_browsers": false
- Custom patterns → provide pattern list →
"in_app_browsers": [...]
Group F — Link Decoration
F1 — Outbound link decoration
Should Intake automatically append UTM parameters and/or click IDs to outbound links (for cross-domain tracking)?
- No (default) → omit
link_decoration
- Yes → ask:
- Which domains to decorate? (comma-separated list)
- Decorate UTM params? (Yes/No)
- Decorate click IDs (gclid, fbclid, etc.)? (Yes/No)
- Any custom params to append?
Save as:
"link_decoration": { "enabled": true, "allowedDomains": [...], "decorateUtm": true/false, "decorateClickIds": true/false }
Group G — Session & Cookie Settings
G1 — Cookie domain
Enter the root domain (e.g. example.com) if cookies need to be shared across subdomains (app.example.com, www.example.com, etc.). Skip if the site is on a single domain.
- If provided:
"domain": "<value>"
- If skipped: omit
G2 — Cookie lifetime
How long should attribution cookies persist?
Default: 6 months. Press Enter to keep default.
- If changed:
"lifetime": <number_of_months>
- If skipped: omit
G3 — Session length
How long should a session last (in minutes) before a new session starts?
Default: 30 minutes. Press Enter to keep default.
- If changed:
"session_length": <minutes>
- If skipped: omit
G4 — Referral restarts session
Should a mid-session referral from an external source start a new session?
Default: No.
- Yes →
"referral_starts_new_session": true
- No (default) → omit
G5 — Timezone offset
Should timestamps be normalized to a specific UTC offset (e.g. 3 for UTC+3)?
Default: visitor's local timezone. Skip to keep default.
- If provided:
"timezone_offset": <number>
- If skipped: omit
Group H — Advanced
H1 — Custom callback
Should Intake call a JavaScript function after initialization (e.g. to send attribution data to your backend)?
- No (default) → omit
- Yes → note the callback function name/expression; add it manually to the generated snippet after generation (generate.py outputs a comment placeholder)
Step 3 — Preview config, ask for confirm
Run the preview:
python3 "$SCRIPT_DIR/generate.py" \
--mode "$MODE" \
--interview /tmp/intake_configurator.json \
--preview
Show the output to the user and ask:
Here is the generated config. Does everything look correct? You can change any parameter or add your own.
- Yes, generate → proceed to Step 4
- Edit → collect changes, update
/tmp/intake_configurator.json, repeat Step 3
Step 4 — Generate output
python3 "$SCRIPT_DIR/generate.py" \
--mode "$MODE" \
--interview /tmp/intake_configurator.json
Output file (written to the current directory):
- Standalone →
intake-snippet.html
- GTM →
intake-gtm-container.json
If the user specified a callback (H1), remind them to add it manually to the generated file.
Step 5 — Show output + deployment instructions
Show the file content in a code block, then give deployment instructions based on mode:
Standalone
Add both tags to the <head> of every page (before </head>):
1. The first tag loads the library from CDN.
2. The second tag initializes it with your settings.
Order matters: the config script must come AFTER the library tag.
GTM
To apply the container:
1. Open Google Tag Manager → select your account and container
2. Admin (gear icon) → Import Container
3. Select the file intake-gtm-container.json
4. Choose workspace → select the target workspace
5. Import type → Merge (not Overwrite!)
6. Click Confirm
7. Verify the created tags in the "Intake" folder:
• "Intake – Library" — loads the library, trigger: All Pages
• "Intake – Config" — initializes with config, fires after Library
8. Preview → confirm both tags fire on All Pages
9. Submit → Version name: "Add Plurio Intake" → Publish
Parameter reference (fallback if GitHub unreachable)
Parameters supported by @plurio/intake:
| Parameter | Type | Default | Interview question |
|---|
domain | string | current domain | G1 |
lifetime | number | 6 months | G2 |
session_length | number | 30 min | G3 |
referral_starts_new_session | boolean | false | G4 |
timezone_offset | number | visitor TZ | G5 |
consent_mode | object | — | A1 |
pii_collection | object | — | A2 |
analytics_ids | object | — | C3 |
data_layer | boolean | true | C2 |
spa_tracking | boolean | true (library) | C1 |
user_id | object | — | B1 |
user_ip | string | — | B2 |
promocode | boolean/object | false | B3 |
referrals | array | 7 social (added by configurator) | D1 |
organics | array | predefined | D2 |
typein_attributes | object | direct/none | D3 |
campaign_param | string | — | D4 |
term_param | string | — | D4 |
content_param | string | — | D4 |
in_app_browsers | array/boolean | enabled | E1 |
link_decoration | object | — | F1 |
callback | function | — | H1 |
Error handling
| Symptom | Cause | Fix |
|---|
python3: command not found | Python not installed | Install Python 3.9+ |
| GitHub fetch fails at Step 0 | No network / repo moved | Use fallback parameter list in this skill |
| GTM "Container JSON is invalid" | File format not accepted | Verify the file contains exportFormatVersion: 2 |
| GTM "Tags already exist" after Merge | Name conflict with existing tags | Rename existing tags in GTM before importing |
intk is not defined in browser | Library tag did not load before Config tag | Verify Library tag is set as Setup Tag for Config tag |
