@growthub/cli 0.3.54 → 0.3.56

package/assets/worker-kits/growthub-twenty-crm-v1/skills.md

@@ -0,0 +1,401 @@
+# Twenty CRM Operator — Master Skill Doc
+
+**Source of truth for methodology. Read this file completely before beginning any task.**
+
+---
+
+## QUICK REFERENCE TABLE
+
+| Resource | Path |
+|---|---|
+| Agent operating law | `workers/twenty-crm-operator/CLAUDE.md` |
+| Brand kit template | `brands/_template/brand-kit.md` |
+| Growthub example brand kit | `brands/growthub/brand-kit.md` |
+| Output contract | `output-standards.md` |
+| Runtime assumptions | `runtime-assumptions.md` |
+| Fork integration notes | `docs/twenty-fork-integration.md` |
+| API and webhooks doc | `docs/api-and-webhooks.md` |
+| Data model layer doc | `docs/data-model-layer.md` |
+| CRM setup brief | `templates/crm-setup-brief.md` |
+| Data model design | `templates/data-model-design.md` |
+| Lead enrichment pipeline | `templates/lead-enrichment-pipeline.md` |
+| Pipeline automation brief | `templates/pipeline-automation-brief.md` |
+| Webhook integration spec | `templates/webhook-integration-spec.md` |
+| API query plan | `templates/api-query-plan.md` |
+| CRM playbook | `templates/crm-playbook.md` |
+| Custom object design | `templates/custom-object-design.md` |
+| Import mapping | `templates/import-mapping.md` |
+| Workspace config checklist | `templates/workspace-config-checklist.md` |
+| Integration handoff | `templates/integration-handoff.md` |
+| Enrichment field map | `templates/enrichment-field-map.md` |
+| Sample CRM setup | `examples/crm-setup-sample.md` |
+| Sample enrichment pipeline | `examples/lead-enrichment-sample.md` |
+| Sample pipeline automation | `examples/pipeline-automation-sample.md` |
+| Sample CRM playbook | `examples/crm-playbook-sample.md` |
+
+---
+
+## STEP 0 — BEFORE ANY TASK, ANSWER THESE QUESTIONS
+
+Before producing anything, confirm:
+
+1. Which client or brand is this for?
+2. What is the primary CRM use case?
+3. What data sources exist today?
+4. What deployment mode is the client using?
+5. Does the client have a local fork available for inspection?
+6. What is the expected go-live scope: data model only / pipeline only / full implementation?
+
+If any of these are unknown after the 3-question gate in CLAUDE.md, stop and ask.
+
+---
+
+## STEP 1 — LOAD THE BRAND KIT
+
+Read `brands/<client-slug>/brand-kit.md` if it exists. Otherwise start from `brands/_template/brand-kit.md`.
+
+Extract:
+- client identity (name, slug, industry, company size)
+- primary CRM objective
+- current tech stack and data sources
+- pipeline stages and deal flow
+- team size and CRM admin capacity
+- target enrichment providers
+- existing deliverables log
+
+The brand kit drives all output naming, object naming conventions, and proposal context.
+
+---
+
+## STEP 2 — CHECK THE WORKING SUBSTRATE
+
+If the user has a local Twenty fork, inspect it before planning anything.
+
+### Source-of-truth file order in the fork
+
+1. `README.md`
+2. `packages/twenty-server/src/modules/` — standard CRM object modules
+3. `packages/twenty-server/src/metadata/` — custom object and field metadata API
+4. `packages/twenty-server/src/api/` — REST and GraphQL endpoint definitions
+5. `packages/twenty-server/src/workflow/` — workflow trigger and action engine
+6. `packages/twenty-front/src/modules/` — UI module definitions
+7. `packages/twenty-cli/src/` — CLI command surface
+8. `docker-compose.yml` — self-hosted stack definition
+
+### What to verify in the fork
+
+- Which standard CRM objects are present (`Person`, `Company`, `Opportunity`, etc.)
+- Whether the metadata API is enabled and accessible
+- What workflow trigger types are implemented
+- Whether webhook outbound actions are supported
+- What the GraphQL schema looks like for the target objects
+- Which CLI commands are available
+- Whether any custom modules are already registered
+
+If the fork cannot be inspected, use the frozen assumptions in `runtime-assumptions.md` and label outputs `assumption-based`.
+
+---
+
+## STEP 3 — WORKFLOW SELECTION LOGIC
+
+Select the narrowest workflow scope that satisfies the real job.
+
+| Workflow | Primary Use | Data Model Required? | Automation Required? |
+|---|---|---|---|
+| `crm-setup` | Net-new workspace from scratch | Yes | Optional |
+| `data-model` | Object schema design only | Yes | No |
+| `lead-enrichment` | Import + enrichment pipeline | Yes | Optional |
+| `pipeline-automation` | Trigger-action automation | Existing model assumed | Yes |
+| `webhook-integration` | External event or outbound hook | Existing model assumed | Yes |
+| `api-reporting` | Query plans for data access | Existing model assumed | No |
+| `full-playbook` | All-in CRM operator documentation | Yes | Yes |
+
+Default selection rules:
+- "Set up Twenty from scratch" → `crm-setup` → escalate to `full-playbook`
+- "Design objects for X use case" → `data-model`
+- "Connect Apollo / Clearbit / Clay" → `lead-enrichment`
+- "Automate follow-up / stage transitions" → `pipeline-automation`
+- "Connect Segment / Stripe / Intercom" → `webhook-integration`
+- "Build a dashboard or export data" → `api-reporting`
+
+---
+
+## STEP 4 — DATA MODEL DESIGN LOGIC
+
+Twenty supports both standard objects and fully custom metadata-defined objects.
+
+### Standard objects (always available)
+
+| Object | Purpose | Key fields |
+|---|---|---|
+| `Person` | Individual contacts | name, email, phone, company, city, linkedInLink, position |
+| `Company` | Organizations | name, domain, employees, annualRecurringRevenue, address |
+| `Opportunity` | Deals and pipeline | name, stage, amount, closeDate, pointOfContact |
+| `Note` | Freeform notes | body, author, target (relation) |
+| `Task` | Action items | title, dueAt, assignee, status, target (relation) |
+| `Workspace Member` | Team users | name, email, role |
+
+### Custom objects
+
+When standard objects do not fit the use case:
+
+1. Define the object in `templates/custom-object-design.md`
+2. Specify all fields (name, type, required, select options if enum)
+3. Define relationships to standard or other custom objects
+4. Document whether creation is via UI (`Settings > Objects`) or metadata API
+5. Define the default view configuration
+
+### Field types available in Twenty
+
+- `TEXT` — single-line text
+- `RICH_TEXT` — multi-line rich text
+- `NUMBER` — numeric (integer or decimal)
+- `BOOLEAN` — true/false toggle
+- `DATE` — date only
+- `DATE_TIME` — date and time
+- `SELECT` — single-choice from predefined options
+- `MULTI_SELECT` — multi-choice from predefined options
+- `RELATION` — link to another object (one-to-one, one-to-many, many-to-many)
+- `LINK` — URL field (label + href)
+- `EMAILS` — multi-email field with primary designation
+- `PHONES` — multi-phone field with primary designation
+- `CURRENCY` — amount + currency code pair
+- `ADDRESS` — structured address (street, city, state, country, postalCode)
+- `RATING` — 1–5 star rating
+
+### Naming conventions
+
+- Object names: PascalCase (`LeadScore`, `PartnerAccount`)
+- Field names: camelCase (`enrichmentSource`, `closeDate`)
+- Select option values: SCREAMING_SNAKE_CASE (`QUALIFIED`, `NOT_INTERESTED`)
+
+---
+
+## STEP 5 — LEAD ENRICHMENT PIPELINE LOGIC
+
+Enrichment pipelines connect external data providers to Twenty CRM contact and company records.
+
+### Supported enrichment patterns
+
+| Pattern | Description |
+|---|---|
+| CSV import + field map | Bulk import from Apollo/Clay export with field-to-Twenty mapping |
+| Webhook ingest | Inbound webhook from enrichment provider populates records |
+| GraphQL mutation | Direct API call to create/update Person or Company records |
+| REST API batch | POST to Twenty REST API with enriched data |
+
+### Deduplication strategy (required)
+
+Every enrichment pipeline must define a match key before import:
+
+| Match key | When to use |
+|---|---|
+| `email` | Primary match for Person objects — highest reliability |
+| `domain` | Primary match for Company objects |
+| `linkedInLink` | Secondary match for Person when email is unavailable |
+| `phone` | Tertiary match — use only when domain and email are both absent |
+| Composite | Email + domain for contacts with company context |
+
+Deduplication behavior in Twenty:
+- On import, Twenty checks for existing records matching the key field
+- If a match is found: update existing record (merge mode)
+- If no match: create new record (insert mode)
+- Always document whether the pipeline runs in merge or insert-only mode
+
+### Enrichment field map requirements
+
+For every provider integration, produce an `EnrichmentFieldMap` that lists:
+- provider field name
+- provider data type
+- target Twenty object
+- target Twenty field name
+- target Twenty field type
+- transformation required (none / normalize / truncate / format)
+- required / optional
+
+Use `templates/enrichment-field-map.md`.
+
+---
+
+## STEP 6 — PIPELINE AUTOMATION LOGIC
+
+Twenty's workflow engine supports trigger-action automation for any object.
+
+### Trigger types
+
+| Trigger | Event |
+|---|---|
+| `on_create` | A record of the specified object type is created |
+| `on_update` | A record is updated (optionally scoped to specific fields) |
+| `on_field_change` | A specific field value changes (supports before/after conditions) |
+| `on_delete` | A record is deleted |
+| `on_schedule` | Time-based trigger (cron expression) |
+| `on_webhook` | Inbound webhook event triggers a workflow |
+
+### Action types
+
+| Action | What it does |
+|---|---|
+| `update_field` | Set one or more fields on the trigger record or a related record |
+| `create_record` | Create a new record of any object type |
+| `send_webhook` | Fire an outbound HTTP POST to an external endpoint |
+| `assign_owner` | Assign the record to a specific workspace member or rotation |
+| `create_task` | Create a Task linked to the trigger record |
+| `send_notification` | In-app notification to a workspace member |
+
+### Automation design rules
+
+- One trigger per workflow — do not chain triggers into a single workflow definition
+- Conditions on `on_update` must specify which field changed and what values are relevant
+- `send_webhook` actions must include the full endpoint, payload schema, and retry policy
+- Every automation must have an explicit "when this fails" path documented
+- Do not design polling loops when `on_update` or `on_webhook` triggers are available
+
+---
+
+## STEP 7 — API AND WEBHOOK INTEGRATION LOGIC
+
+### GraphQL API
+
+Twenty's GraphQL API is the primary data access surface.
+
+Common query patterns:
+
+```graphql
+# Fetch all persons with company relation
+query {
+  people {
+    edges {
+      node {
+        id
+        name { firstName lastName }
+        emails { primaryEmail }
+        company { name domain }
+        createdAt
+      }
+    }
+  }
+}
+
+# Create a person record
+mutation {
+  createPerson(data: {
+    name: { firstName: "Jane", lastName: "Doe" }
+    emails: { primaryEmail: "jane@example.com" }
+    companyId: "<company-id>"
+  }) {
+    id
+    createdAt
+  }
+}
+
+# Update an opportunity stage
+mutation {
+  updateOpportunity(
+    id: "<opportunity-id>"
+    data: { stage: "CLOSING" }
+  ) {
+    id
+    stage
+    updatedAt
+  }
+}
+```
+
+### REST API
+
+Twenty provides a REST API at `/api/objects/<object-name>` for simple CRUD operations.
+
+```bash
+# GET all companies
+GET /api/objects/companies
+
+# POST a new person
+POST /api/objects/people
+Content-Type: application/json
+Authorization: Bearer <app-token>
+
+# PATCH update a field
+PATCH /api/objects/opportunities/<id>
+```
+
+### Webhooks
+
+Twenty emits webhook events for object lifecycle events.
+
+Event naming pattern: `<object>.<event>` (e.g., `opportunity.created`, `person.updated`)
+
+Every webhook spec must define:
+- event name
+- trigger object and conditions
+- payload schema (what Twenty sends)
+- target endpoint URL pattern
+- auth mechanism on the receiving end
+- retry policy (Twenty default: 3 attempts with exponential backoff)
+
+---
+
+## STEP 8 — IMPORT AND MIGRATION LOGIC
+
+When migrating from another CRM (Salesforce, HubSpot, Pipedrive, Airtable):
+
+### Import workflow
+
+1. Export data from source CRM as CSV
+2. Clean and deduplicate the CSV (remove test records, normalize email format)
+3. Produce an `ImportMapping` document that maps every source column to a Twenty field
+4. Run a sample import (first 50 rows) and verify field alignment
+5. Run the full import
+6. Run a post-import audit: record count, duplicate check, relation integrity
+
+### Common source-to-Twenty field mappings
+
+| Source (generic) | Twenty Field | Object | Notes |
+|---|---|---|---|
+| First Name | `name.firstName` | Person | |
+| Last Name | `name.lastName` | Person | |
+| Email | `emails.primaryEmail` | Person | Use as dedup key |
+| Phone | `phones.primaryPhoneNumber` | Person | |
+| Company Name | `company.name` | Company | Create or link |
+| Company Domain | `company.domain` | Company | Use as dedup key |
+| Job Title | `position` | Person | |
+| LinkedIn URL | `linkedInLink.url` | Person | |
+| Deal Name | `name` | Opportunity | |
+| Deal Stage | `stage` | Opportunity | Map to Twenty select values |
+| Deal Value | `amount.amountMicros` | Opportunity | Multiply by 1,000,000 for micros |
+| Close Date | `closeDate` | Opportunity | ISO 8601 format |
+
+---
+
+## STEP 9 — OUTPUT ORDER
+
+Produce artifacts in this strict order:
+
+1. CRM Setup Brief (`templates/crm-setup-brief.md`)
+2. Data Model Design (`templates/data-model-design.md`)
+3. Lead Enrichment Pipeline (`templates/lead-enrichment-pipeline.md`)
+4. Pipeline Automation Brief (`templates/pipeline-automation-brief.md`)
+5. Webhook Integration Spec (`templates/webhook-integration-spec.md`)
+6. API Query Plan (`templates/api-query-plan.md`)
+7. Custom Object Design (`templates/custom-object-design.md`) — if custom objects required
+8. Import Mapping (`templates/import-mapping.md`) — if data migration required
+9. Workspace Config Checklist (`templates/workspace-config-checklist.md`)
+10. Integration Handoff (`templates/integration-handoff.md`)
+11. Enrichment Field Map (`templates/enrichment-field-map.md`)
+12. CRM Playbook (`templates/crm-playbook.md`)
+
+---
+
+## STEP 10 — QUALITY BAR
+
+Good output looks like this:
+
+- all object schemas are grounded in Twenty's actual field types and naming conventions
+- data model is complete before any automation or pipeline spec is written
+- deduplication key is explicitly named in every import or enrichment pipeline
+- every automation trigger-action pair has a named failure path
+- webhook specs include the full payload schema, not just the event name
+- GraphQL queries are syntactically correct and reference real Twenty field paths
+- CRM playbook is written at a level a non-technical team member can execute
+- every output file can be handed to a developer or CRM admin and acted on immediately
+- no filler paragraphs — every sentence either presents a decision, explains a constraint, or specifies an action

package/assets/worker-kits/growthub-twenty-crm-v1/templates/api-query-plan.md

@@ -0,0 +1,179 @@
+# API Query Plan — [CLIENT NAME]
+
+**Date:** YYYY-MM-DD  
+**Kit:** `growthub-twenty-crm-v1`  
+**Mode:** `[cloud / self-hosted / local-fork / agent-only]`
+
+---
+
+## QUERY INVENTORY
+
+| # | Name | Type | Object | Use case |
+|---|---|---|---|---|
+| 1 | [Name] | [GraphQL / REST] | [object] | [use case] |
+| 2 | [Name] | [GraphQL / REST] | [object] | [use case] |
+
+---
+
+## QUERY 1 — [NAME]
+
+**Use case:** [What this query supports — e.g. "Weekly pipeline report: all open Opportunities by stage"]  
+**Consumer:** [dashboard / export script / integration / reporting tool]  
+**Type:** `GraphQL`
+
+### Query
+
+```graphql
+query GetOpenOpportunitiesByStage($filter: OpportunityFilterInput) {
+  opportunities(filter: $filter, orderBy: { createdAt: DescNullsLast }) {
+    edges {
+      node {
+        id
+        name
+        stage
+        amount { amountMicros currencyCode }
+        closeDate
+        pointOfContact {
+          name { firstName lastName }
+          emails { primaryEmail }
+        }
+        company { name domain }
+        createdAt
+        updatedAt
+      }
+    }
+    totalCount
+  }
+}
+```
+
+**Variables:**
+
+```json
+{
+  "filter": {
+    "stage": { "in": ["LEAD", "QUALIFIED", "DEMO", "PROPOSAL"] }
+  }
+}
+```
+
+**Expected response structure:**
+
+```json
+{
+  "data": {
+    "opportunities": {
+      "edges": [
+        {
+          "node": {
+            "id": "...",
+            "name": "...",
+            "stage": "QUALIFIED",
+            "amount": { "amountMicros": 10000000, "currencyCode": "USD" }
+          }
+        }
+      ],
+      "totalCount": 42
+    }
+  }
+}
+```
+
+**Authentication:**
+
+```
+Authorization: Bearer <TWENTY_API_TOKEN>
+Content-Type: application/json
+POST <TWENTY_API_URL>/graphql
+```
+
+---
+
+## QUERY 2 — [NAME]
+
+**Use case:** [description]  
+**Consumer:** [consumer]  
+**Type:** `[GraphQL / REST]`
+
+### REST Query (if applicable)
+
+```bash
+GET <TWENTY_API_URL>/api/objects/people?filter=emails.primaryEmail[eq]=jane@example.com
+Authorization: Bearer <TWENTY_API_TOKEN>
+```
+
+### GraphQL Query (if applicable)
+
+```graphql
+query {
+  # [query here]
+}
+```
+
+---
+
+## QUERY 3 — MUTATION: CREATE PERSON
+
+**Use case:** [e.g. "Create a new Person record from enrichment pipeline"]  
+**Type:** `GraphQL mutation`
+
+```graphql
+mutation CreatePerson($data: PersonCreateInput!) {
+  createPerson(data: $data) {
+    id
+    name { firstName lastName }
+    emails { primaryEmail }
+    createdAt
+  }
+}
+```
+
+**Variables:**
+
+```json
+{
+  "data": {
+    "name": { "firstName": "Jane", "lastName": "Doe" },
+    "emails": { "primaryEmail": "jane@example.com" },
+    "position": "Head of Growth",
+    "linkedInLink": { "url": "https://linkedin.com/in/janedoe", "label": "LinkedIn" }
+  }
+}
+```
+
+---
+
+## QUERY 4 — MUTATION: UPDATE OPPORTUNITY STAGE
+
+**Use case:** [e.g. "Automation updates stage when deal is marked as won"]  
+**Type:** `GraphQL mutation`
+
+```graphql
+mutation UpdateOpportunityStage($id: ID!, $stage: String!) {
+  updateOpportunity(id: $id, data: { stage: $stage }) {
+    id
+    stage
+    updatedAt
+  }
+}
+```
+
+**Variables:**
+
+```json
+{
+  "id": "<opportunity-id>",
+  "stage": "CLOSED_WON"
+}
+```
+
+---
+
+## RATE LIMITS AND CONSTRAINTS
+
+| Constraint | Value | Notes |
+|---|---|---|
+| GraphQL endpoint | `<TWENTY_API_URL>/graphql` | All mutations and queries |
+| REST endpoint | `<TWENTY_API_URL>/api/objects/<object>` | CRUD only |
+| Auth | Bearer token | Workspace-scoped |
+| [Rate limit if known] | [value] | [notes] |