PRD-020: VibesPro™ Foundation Platform Requirements

Status: Proposed Version: 1.0 Date: 2025-12-27 Implements ADR: ADR-030 (VibesPro™ Foundation Integration Strategy) Related PRDs: PRD-004 (SEA™ Forge Delivery Pipeline)

1. Executive Summary

This PRD defines the product requirements for adopting VibesPro™’s core systems as the foundational platform for SEA-Forge™. It translates ADR-030’s technical decisions into concrete features, user stories, and acceptance criteria.

2. Problem Statement

Current State

SEA™ has spec-based generation but lacks implemented runtime infrastructure
Temporal Database Service (SDS-015) is specified but not implemented
Type synchronization is documented but not automated
Developer experience is fragmented across multiple tools

Desired State

VibesPro™ provides proven implementations for all core subsystems
Unified developer experience across spec authoring and code generation
Institutional memory via Temporal AI informs all code generation decisions
End-to-end type safety from database to frontend

3. Goals & Non-Goals

Goals

G-001: Adopt VibesPro™ Temporal AI as SEA™ Pattern Oracle
G-002: Implement end-to-end type safety pipeline
G-003: Inherit VibesPro™’s 5 user journeys for SEA™ developers
G-004: Integrate meta-generator system for “No-Agent-Writes-Code”
G-005: Adopt Context-Kit for SEA™ agent orchestration

Non-Goals

Forking or modifying VibesPro™ core
Replacing SEA™ DSL with VibesPro™-native language
Changing SEA™ governance model

4. User Stories

Epic 1: Temporal AI Integration

US-001: Pattern Query During Generation

As a SEA™ generator author
I want to query historical patterns before generating code
So that I can leverage institutional knowledge and avoid anti-patterns

Acceptance Criteria:

just temporal-ai-query "hexagonal repository" returns ranked patterns
Pattern recommendations include success rates from production metrics
Results are filterable by domain, language, and governance status

US-002: Automatic Pattern Learning

As a SEA™ platform maintainer
I want generated code patterns to be automatically embedded in Temporal AI
So that the pattern library grows with each generation

Acceptance Criteria:

Post-generation hook stores embeddings in Temporal DB
Embeddings include traceability to source specs (SDS-xxx)
Refresh mechanism updates patterns from Git history

Epic 2: End-to-End Type Safety

US-003: Database-to-TypeScript Generation

As a SEA™ developer
I want to generate TypeScript types from my Supabase schema
So that my frontend is always type-safe against the database

Acceptance Criteria:

just gen-types-ts generates database.types.ts
Types include all tables, views, and functions
Breaking schema changes fail CI before merge

US-004: TypeScript-to-Pydantic Transpilation

As a SEA™ backend developer
I want Python Pydantic models generated from TypeScript types
So that I have runtime validation matching my database schema

Acceptance Criteria:

just gen-types-py generates Pydantic models
Models include validators matching TypeScript constraints
Bidirectional consistency is enforced in CI

Epic 3: Developer Experience Journeys

US-005: Journey A - Project Generation

As a new SEA™ developer
I want to generate a complete project with all SEA™ governance
So that I can start building immediately with best practices

Acceptance Criteria:

copier copy gh:GodSpeedAI/SEA™ my-project generates working project
Generated project includes SEA™ DSL validation hooks
Generated project includes IFL identity minting
just setup succeeds with all dependencies installed

US-006: Journey C - AI-Assisted Development

As an AI-assisted SEA™ developer
I want to use GitHub Copilot with SEA-aware context
So that generated code follows SEA™ governance

Acceptance Criteria:

.github/instructions/*.md files provide SEA™ context
ce.manifest.jsonc includes SEA-specific skills
AI suggestions respect hexagonal architecture boundaries

Epic 4: Meta-Generator System

US-007: Generator Creation

As a SEA™ platform developer
I want to create new generators using a meta-generator
So that I can extend the generation system without boilerplate

Acceptance Criteria:

just generator-new my-generator type=domain creates generator scaffold
Generated generator includes schema.json, tests, and spec documentation
Generator integrates with SEA™ validation pipeline

US-008: Hexagonal Code Generation

As a SEA™ domain modeler
I want to generate hexagonal architecture code from specs
So that my domain models follow clean architecture

Acceptance Criteria:

Generator produces domain/, application/, infrastructure/ directories
Ports and adapters are correctly separated
Generated code includes spec traceability comments

Epic 5: Observability Integration

US-009: Unified Telemetry

As a SEA™ operations engineer
I want all SEA™ services to emit OpenTelemetry spans
So that I can trace requests across the semantic pipeline

Acceptance Criteria:

Vector configuration ingests OTLP from all services
Spans are correlated with SEA™ DSL node identifiers
PII is redacted before export to OpenObserve

5. Functional Requirements

FR-001: Temporal AI CLI Integration

Requirement: SEA™ MUST expose Temporal AI via just temporal-ai-* commands
Justification: Enables pattern-assisted code generation
Priority: P0

FR-002: Type Generation Pipeline

Requirement: SEA™ MUST implement just gen-types with TS and Python targets
Justification: Enforces isomorphic typing across stack
Priority: P0

FR-003: Copier Template Extension

Requirement: SEA™ MUST extend VibesPro™ Copier templates with governance hooks
Justification: Enables governed project generation
Priority: P0

FR-004: Context-Kit Manifest

Requirement: SEA™ MUST maintain ce.manifest.jsonc with SEA™ skills
Justification: Enables AI routing with SEA™ context
Priority: P1

FR-005: Meta-Generator Integration

Requirement: SEA™ MUST expose generator creation via Nx
Justification: Enables extensible generation system
Priority: P1

6. Non-Functional Requirements

NFR-001: Performance

Metric	Target	Measurement
Generator execution	< 5s	Wall clock time per generator
Temporal AI query	< 100ms	P95 latency
Type generation	< 30s	Full TS + Python pipeline
CI validation	< 5min	Complete validation suite

NFR-002: Reliability

Generation MUST be idempotent
Type sync MUST be deterministic (same schema → same types)
Temporal AI MUST degrade gracefully if unavailable

NFR-003: Security

Secrets MUST be managed via SOPS
Generated code MUST NOT contain credentials
Temporal AI embeddings MUST NOT contain PII

7. Dependencies

Dependency	Version	Purpose
VibesPro™	main branch	Foundation substrate
Temporal AI	crates/temporal-ai	Pattern oracle
Context-Kit	ce.manifest.jsonc	Agent routing
Copier	9.x	Template engine
Nx	21.x	Monorepo orchestration

8. Success Metrics

Metric	Target	Baseline
Developer onboarding time	< 30 minutes	N/A (new)
Pattern recommendation adoption	> 60%	N/A (new)
Type sync violations in CI	0	N/A (new)
Generator test coverage	> 80%	N/A (new)

9. Risks & Mitigations

Risk	Likelihood	Impact	Mitigation
VibesPro™ breaking changes	Medium	High	Pin to stable releases, maintain fork capability
Temporal AI performance	Low	Medium	Add caching layer, benchmark continuously
Type generation mismatch	Medium	High	CI gates with bidirectional validation
Developer learning curve	Medium	Medium	Comprehensive documentation, tutorials

10. Implementation Phases

Phase 1: Foundation (Week 1-2)

Clone VibesPro™ structure into SEA™
Extend ce.manifest.jsonc with SEA™ skills
Add SEA-specific Copier questions

Phase 2: Temporal AI (Week 3-4)

Deploy Temporal AI in SEA™
Integrate with generator pipeline
Add pattern recommendation UI

Phase 3: Type Safety (Week 5-6)

Implement type generation pipeline
Add CI validation gates
Document type authoring workflow

Phase 4: Observability (Week 7-8)

Align observability stack
Add semantic correlation
Deploy monitoring dashboards