Generate Boundary Objects

You are generating strategic boundary objects for a studio project. This creates the strategic context needed for targeted experiment generation.

Input

The user has provided: $ARGUMENTS

This should be a studio project slug (e.g., chalk, putt).

JSONB Block Format

All canvas JSONB block columns (jobs, pains, gains, products_services, pain_relievers, gain_creators, key_partners, etc.) use this structure:

json
1{
2  "items": [
3    {
4      "id": "<unique-id>",
5      "content": "<item text>",
6      "priority": "high|medium|low",
7      "created_at": "<ISO 8601 timestamp>"
8    }
9  ],
10  "assumptions": [],
11  "validation_status": "untested"
12}

For customer profile items, also include metadata:

json
1{
2  "id": "<unique-id>",
3  "content": "<item text>",
4  "metadata": { "type": "functional|social|emotional" },
5  "priority": "high|medium|low",
6  "created_at": "<ISO 8601 timestamp>"
7}

Generate unique IDs using a prefix + sequence pattern (e.g., job-001, pain-001, ps-001). Use new Date().toISOString() for timestamps.

CRITICAL: Do NOT store plain string arrays. The admin UI canvas views will not render them.

Required Fields by Table

Schema source of truth: lib/mcp/schemas/*.ts If any field documentation here conflicts with the Zod schemas, the schemas are correct. Last synced: 2026-02-21

business_model_canvases

Required: slug, name, status Optional: description, studio_project_id, all 9 block columns (JSONB format above)

customer_profiles

Required: slug, name, status Optional: description, studio_project_id, profile_type, demographics (JSONB object), jobs (JSONB), pains (JSONB), gains (JSONB)

value_maps

Required: slug, name, status Optional: description, studio_project_id, customer_profile_id, business_model_canvas_id, products_services (JSONB), pain_relievers (JSONB), gain_creators (JSONB)

value_proposition_canvases

Required: slug, name, status, value_map_id Optional: description, studio_project_id, customer_profile_id, addressed_jobs, addressed_pains, addressed_gains

assumptions

Required: slug, statement, status (use identified), importance (critical|high|medium|low) Optional: studio_project_id, category, notes, validation_criteria Note: is_leap_of_faith is computed by DB trigger — do not provide

user_journeys

Required: slug, name, status Optional: description, studio_project_id, customer_profile_id, journey_type (use end_to_end), goal

journey_stages

Required: user_journey_id, name, sequence Optional: description, stage_type (pre_purchase|purchase|post_purchase|ongoing), customer_emotion, customer_mindset, customer_goal, drop_off_risk (low|medium|high)

touchpoints

Required: journey_stage_id, name, sequence Optional: description, channel_type (web|mobile_app|phone|email|in_person|chat|social|physical_location|mail|other), interaction_type (browse|search|read|watch|listen|form|transaction|conversation|notification|passive), importance (critical|high|medium|low), pain_level (none|minor|moderate|major|critical), delight_potential (low|medium|high)

service_blueprints

Required: slug, name, status Optional: description, studio_project_id, blueprint_type (service|product|hybrid|digital|physical), service_scope, service_duration

blueprint_steps

Required: service_blueprint_id, name, sequence Optional: description, layers (JSONB with keys: customer_action, frontstage, backstage, support_process), cost_implication (none|low|medium|high), customer_value_delivery (none|low|medium|high), failure_risk (low|medium|high|critical), failure_impact

story_maps

Required: slug, name, status Optional: description, studio_project_id, map_type (feature|product|release|discovery)

activities

Required: story_map_id, name, sequence Optional: description, user_goal

user_stories

Required: activity_id, title Optional: description, acceptance_criteria, story_type (feature|enhancement|bug|tech_debt|spike), priority (critical|high|medium|low), status (default: backlog), story_points

entity_links

Required: source_type, source_id, target_type, target_id Optional: link_type (default: 'related'), strength (strong|moderate|weak|tentative), notes, metadata, position

Procedure

Step 1: Validate Project

bash
1scripts/sb query studio_projects "slug=eq.<slug>"

If not found, abort with: "No studio project found with slug '<slug>'. Check studio_projects table."

Save the project id, name, description, problem_statement, hypothesis, success_criteria, current_focus, scope_out.

Step 2: Check Existing Links

Query entity_links for existing outbound links from this project:

bash
1scripts/sb query entity_links "source_type=eq.studio_project&source_id=eq.<project-id>"

Show what's already linked. Skip types that already have links to keep this idempotent. Report which types are being skipped and why.

Step 3: Read Project Context

Also query studio_hypotheses for this project:

bash
1scripts/sb query studio_hypotheses "project_id=eq.<project-id>"

Gather all available context: name, description, problem_statement, hypothesis, success_criteria, current_focus, scope_out, plus hypothesis statements.

Step 4: Generate Business Model Canvas

Skip if an entity_link to business_model_canvas already exists.

Create a business_model_canvases record. Populate ALL 9 blocks with initial items derived from project context:

bash
1scripts/sb create business_model_canvases '{
2  "slug": "<project-slug>-business-model",
3  "name": "<Project Name> - Business Model",
4  "description": "<1-2 sentences grounded in project context>",
5  "studio_project_id": "<project-id>",
6  "status": "draft",
7  "key_partners": {"items": [{"id": "kp-001", "content": "<partner>", "priority": "high", "created_at": "<now>"}], "assumptions": [], "validation_status": "untested"},
8  "key_activities": {"items": [...], "assumptions": [], "validation_status": "untested"},
9  "key_resources": {"items": [...], "assumptions": [], "validation_status": "untested"},
10  "value_propositions": {"items": [...], "assumptions": [], "validation_status": "untested"},
11  "customer_segments": {"items": [...], "assumptions": [], "validation_status": "untested"},
12  "customer_relationships": {"items": [...], "assumptions": [], "validation_status": "untested"},
13  "channels": {"items": [...], "assumptions": [], "validation_status": "untested"},
14  "cost_structure": {"items": [...], "assumptions": [], "validation_status": "untested"},
15  "revenue_streams": {"items": [...], "assumptions": [], "validation_status": "untested"}
16}'

Generate 2-4 items per block, grounded in the project's problem statement and description. Then create the entity_link:

bash
1scripts/sb create entity_links '{
2  "source_type": "studio_project",
3  "source_id": "<project-id>",
4  "target_type": "business_model_canvas",
5  "target_id": "<bmc-id>",
6  "link_type": "explores",
7  "metadata": {}
8}'

Step 5: Generate Customer Profile

Skip if an entity_link to customer_profile already exists.

Create a customer_profiles record. Use the JSONB block format for jobs, pains, and gains:

bash
1scripts/sb create customer_profiles '{
2  "slug": "<project-slug>-primary-customer",
3  "name": "<Project Name> - Primary Customer",
4  "description": "<2-3 sentences describing the target customer>",
5  "studio_project_id": "<project-id>",
6  "profile_type": "persona",
7  "demographics": {"age_range": "<Age range>", "role": "<Role>", "tech_level": "<level>", "income": "<range>"},
8  "jobs": {"items": [...], "assumptions": [], "validation_status": "untested"},
9  "pains": {"items": [...], "assumptions": [], "validation_status": "untested"},
10  "gains": {"items": [...], "assumptions": [], "validation_status": "untested"},
11  "status": "draft"
12}'

Generate 3-5 items per block. Link with explores.

Step 6: Generate Value Map + Value Proposition Canvas

Skip if an entity_link to value_proposition_canvas already exists.

The VPC requires a value_map_id, so create the value map first:

bash
1scripts/sb create value_maps '{
2  "slug": "<project-slug>-value-map",
3  "name": "<Project Name> - Value Map",
4  "description": "<1-2 sentences>",
5  "studio_project_id": "<project-id>",
6  "customer_profile_id": "<profile-id>",
7  "business_model_canvas_id": "<bmc-id>",
8  "products_services": {"items": [...], "assumptions": [], "validation_status": "untested"},
9  "pain_relievers": {"items": [...], "assumptions": [], "validation_status": "untested"},
10  "gain_creators": {"items": [...], "assumptions": [], "validation_status": "untested"},
11  "status": "draft"
12}'

Then create the VPC linked to the value map:

bash
1scripts/sb create value_proposition_canvases '{
2  "slug": "<project-slug>-value-proposition",
3  "name": "<Project Name> - Value Proposition",
4  "description": "<Value prop description>",
5  "studio_project_id": "<project-id>",
6  "customer_profile_id": "<profile-id>",
7  "value_map_id": "<value-map-id>",
8  "addressed_jobs": {"items": ["<job content>", "<job content>"]},
9  "addressed_pains": {"items": ["<pain content>", "<pain content>"]},
10  "addressed_gains": {"items": ["<gain content>", "<gain content>"]},
11  "status": "draft"
12}'

Link the VPC with explores.

Step 7: Generate Assumptions

Skip if entity_links to assumption already exist (3+ links).

Create 3-5 assumptions records. Distribute across risk categories (desirability, viability, feasibility, usability). Bulk insert as an array:

bash
1scripts/sb create assumptions '[
2  {
3    "slug": "<project-slug>-<short-name>",
4    "statement": "We assume that <specific belief>",
5    "status": "identified",
6    "importance": "critical",
7    "category": "desirability",
8    "studio_project_id": "<project-id>",
9    "notes": "<Why this matters>"
10  },
11  { ... },
12  { ... }
13]'

Then bulk-create entity_links for all assumptions:

bash
1scripts/sb create entity_links '[
2  {"source_type": "studio_project", "source_id": "<project-id>", "target_type": "assumption", "target_id": "<id-1>", "link_type": "tests", "metadata": {}},
3  {"source_type": "studio_project", "source_id": "<project-id>", "target_type": "assumption", "target_id": "<id-2>", "link_type": "tests", "metadata": {}},
4  ...
5]'

Step 8: Generate User Journey + Stages + Touchpoints

Skip if an entity_link to user_journey already exists.

Create the full journey cascade — journey record, stages, and touchpoints:

8a. Create the journey:

bash
1scripts/sb create user_journeys '{
2  "slug": "<project-slug>-core-journey",
3  "name": "<Project Name> - Core Journey",
4  "description": "<Journey description grounded in project context>",
5  "studio_project_id": "<project-id>",
6  "customer_profile_id": "<profile-id>",
7  "journey_type": "end_to_end",
8  "goal": "<What success looks like for this journey>",
9  "status": "draft"
10}'

8b. Bulk-create 4-5 journey stages:

bash
1scripts/sb create journey_stages '[
2  {"user_journey_id": "<journey-id>", "name": "Discovery", "description": "...", "sequence": 1, "stage_type": "pre_purchase", "customer_emotion": "...", "customer_mindset": "...", "customer_goal": "...", "drop_off_risk": "medium"},
3  {"user_journey_id": "<journey-id>", "name": "Evaluation", "description": "...", "sequence": 2, "stage_type": "pre_purchase", ...},
4  {"user_journey_id": "<journey-id>", "name": "Commitment", "description": "...", "sequence": 3, "stage_type": "purchase", ...},
5  {"user_journey_id": "<journey-id>", "name": "Onboarding", "description": "...", "sequence": 4, "stage_type": "post_purchase", ...},
6  {"user_journey_id": "<journey-id>", "name": "Ongoing Usage", "description": "...", "sequence": 5, "stage_type": "ongoing", ...}
7]'

Adapt stage names to match the project domain.

8c. Bulk-create touchpoints per stage (2-3 per stage):

bash
1scripts/sb create touchpoints '[
2  {"journey_stage_id": "<stage-1-id>", "name": "...", "sequence": 1, "channel_type": "web", "interaction_type": "browse", "importance": "high", "pain_level": "minor", "delight_potential": "medium"},
3  {"journey_stage_id": "<stage-1-id>", "name": "...", "sequence": 2, ...},
4  {"journey_stage_id": "<stage-2-id>", "name": "...", "sequence": 1, ...},
5  ...
6]'

8d. Create entity_link:

bash
1scripts/sb create entity_links '{
2  "source_type": "studio_project",
3  "source_id": "<project-id>",
4  "target_type": "user_journey",
5  "target_id": "<journey-id>",
6  "link_type": "explores",
7  "metadata": {}
8}'

Step 9: Generate Service Blueprint + Steps

Skip if an entity_link to service_blueprint already exists.

The service blueprint mirrors the journey stages, mapping what happens behind the scenes.

9a. Create the blueprint:

bash
1scripts/sb create service_blueprints '{
2  "slug": "<project-slug>-service-blueprint",
3  "name": "<Project Name> - Service Blueprint",
4  "description": "<How the service is delivered>",
5  "studio_project_id": "<project-id>",
6  "blueprint_type": "digital",
7  "status": "draft"
8}'

9b. Bulk-create blueprint_steps (one per journey stage):

bash
1scripts/sb create blueprint_steps '[
2  {"service_blueprint_id": "<blueprint-id>", "name": "Discovery", "description": "...", "sequence": 1, "layers": {"customer_action": "...", "frontstage": "...", "backstage": "...", "support_process": "..."}, "cost_implication": "low", "customer_value_delivery": "medium", "failure_risk": "low"},
3  {"service_blueprint_id": "<blueprint-id>", "name": "Evaluation", ...},
4  ...
5]'

Populate all 4 layer keys with project-specific content derived from the journey stages and project context.

9c. Create entity_links:

bash
1scripts/sb create entity_links '[
2  {"source_type": "studio_project", "source_id": "<project-id>", "target_type": "service_blueprint", "target_id": "<blueprint-id>", "link_type": "prototypes", "metadata": {}},
3  {"source_type": "service_blueprint", "source_id": "<blueprint-id>", "target_type": "user_journey", "target_id": "<journey-id>", "link_type": "implements", "metadata": {}}
4]'

Step 10: Generate Story Map + Activities + Stories

Skip if an entity_link to story_map already exists.

The story map translates the journey into development work.

10a. Create the story map:

bash
1scripts/sb create story_maps '{
2  "slug": "<project-slug>-story-map",
3  "name": "<Project Name> - Story Map",
4  "description": "<Dev planning for the project>",
5  "studio_project_id": "<project-id>",
6  "map_type": "feature",
7  "status": "draft"
8}'

10b. Bulk-create activities (one per journey stage):

bash
1scripts/sb create activities '[
2  {"story_map_id": "<story-map-id>", "name": "Discovery", "description": "...", "sequence": 1, "user_goal": "..."},
3  {"story_map_id": "<story-map-id>", "name": "Evaluation", "description": "...", "sequence": 2, "user_goal": "..."},
4  ...
5]'

10c. Bulk-create user stories (2-3 per activity):

bash
1scripts/sb create user_stories '[
2  {"activity_id": "<activity-1-id>", "title": "As a <persona>, I want to <action> so that <benefit>", "description": "...", "story_type": "feature", "priority": "high", "status": "backlog"},
3  {"activity_id": "<activity-1-id>", "title": "As a ...", ...},
4  {"activity_id": "<activity-2-id>", "title": "As a ...", ...},
5  ...
6]'

Use the user story format: "As a [persona], I want to [action] so that [benefit]". Ground stories in the customer profile and journey touchpoints.

10d. Create entity_links:

bash
1scripts/sb create entity_links '[
2  {"source_type": "studio_project", "source_id": "<project-id>", "target_type": "story_map", "target_id": "<story-map-id>", "link_type": "informs", "metadata": {}},
3  {"source_type": "story_map", "source_id": "<story-map-id>", "target_type": "user_journey", "target_id": "<journey-id>", "link_type": "implements", "metadata": {}},
4  {"source_type": "story_map", "source_id": "<story-map-id>", "target_type": "service_blueprint", "target_id": "<blueprint-id>", "link_type": "implements", "metadata": {}}
5]'

Step 11: Summary

Present a summary of everything created:

## Boundary Objects Generated for <Name>

### Created:
- Business Model Canvas: "<name>" (id: <id>) — 9 blocks populated
- Customer Profile: "<name>" (id: <id>) — jobs/pains/gains populated
- Value Map: "<name>" (id: <id>) — products/relievers/creators populated
- Value Proposition Canvas: "<name>" (id: <id>)
- Assumptions: <count> created
  - <statement summary> (<category>, <importance>)
  - ...
- User Journey: "<name>" (id: <id>) — <N> stages, <M> touchpoints
- Service Blueprint: "<name>" (id: <id>) — <N> steps with 4-layer structure
- Story Map: "<name>" (id: <id>) — <N> activities, <M> user stories

### Cascade:
Journey → Blueprint → Story Map (all cross-linked via entity_links)

### Skipped (already linked):
- <list any types that were skipped>

### Next Steps:
1. Review boundary objects in admin UI
2. Run `/generate-experiments <slug>` to create context-aware experiments
3. Refine BMC blocks and canvas items as understanding deepens

Error Handling

API Failures

If any scripts/sb call fails:

1. Report the error: Show the exact error message and which step failed
2. Don't continue blindly: If a dependency fails (e.g., BMC creation fails), skip steps that depend on it (e.g., don't create entity_link for BMC)
3. Partial success is OK: If 3 of 5 assumptions succeed, report what worked and what didn't
4. Common errors:
   • duplicate key value violates unique constraint — a record with that slug already exists. Suggest a different slug or offer to link the existing record instead.
   • violates check constraint — a field value is invalid. Check the "Required Fields by Table" section for valid enum values.
   • violates foreign key constraint — a referenced record doesn't exist. Verify the ID is correct.

Reporting

After completion, always present a clear summary showing:

• What was created successfully (with IDs)
• What was skipped (already existed)
• What failed (with error details and suggested fixes)

Important Notes

• Idempotent: Always check existing links before creating. Never duplicate boundary objects.
• Use scripts/sb: All database operations go through scripts/sb commands via the Bash tool.
• Bulk inserts: Use array JSON for journey_stages, touchpoints, blueprint_steps, activities, user_stories, assumptions, and entity_links. This is one of the key wins — a single POST instead of N sequential calls.
• Draft status: All generated objects start as draft — they need human review.
• JSONB block format: All canvas block columns MUST use { items: [...], assumptions: [], validation_status: "untested" } format. Plain arrays will not render in the admin UI.
• Context-grounded: Every generated object should be grounded in the project's actual problem_statement, hypothesis, and description. Don't generate generic content.
• Entity link validation: The studio_project source type is registered in lib/entity-links-validation.ts with valid link types: explores, improves, prototypes, informs, tests, validates, spin_off, inspired_by, evolved_from.
• Required fields: Every table requires slug. Assumptions use status: 'identified' (not untested). User journeys use journey_type: 'end_to_end'. VPCs require value_map_id.