Overview

Specialty Layers provide domain-specific configurations that can be loaded into the agent framework. Each layer defines a complete agent profile: system prompts, tools, RAG collections, context adapters, project templates, and a questionnaire for customizing behavior. Layers are discovered automatically from the filesystem — no manual registration needed.

Built-in Layers (12)

blazor_developer

Blazor web application developer

clean_code_rules

Clean code validation and enforcement rules

django_developer

Django web framework developer

domain_architecture_rules

Domain-driven design architecture rules

fastapi_developer

FastAPI web framework developer

flask_developer

Flask web framework developer

maui_developer

.NET MAUI cross-platform developer

react_developer

React frontend developer

vue_developer

Vue.js frontend developer

web_designer

Web design and UX specialist

winforms_designer

Windows Forms UI designer

wpf_developer

WPF desktop application developer

Core Services (app/services/specialty_layers/)

ModuleKey ClassResponsibility
contracts.pySpecialtyLayerManifest, QuestionnaireTyped data contracts and validation
registry.pySpecialtyLayerDiscovery, SpecialtyLayerRegistryFilesystem discovery, load, enable/disable, list
lifecycle_service.pySpecialtyLayerLifecycleServiceOrchestrates all bindings + rules injection
questionnaire_engine.pySpecialtyLayerQuestionnaireEngineRender, validate, state machine for questionnaires
rules_layer.pyRulesLayerInjector, RulesLayerValidatorRules injection and validation
tool_binding.pySpecialtyLayerToolBindingRegister layer tools with namespacing
rag_binding.pySpecialtyLayerRAGBindingWire layer RAG collections
adapter_binding.pySpecialtyLayerContextAdapterBindingWire context adapters
template_binding.pySpecialtyLayerTemplateBindingWire project templates
service_metadata.pyService metadataService info, setup, status for UI

Layer File Structure

specialty_layers/<layer_name>/
├── layer.yaml          # Manifest: id, name, type, domain, questionnaire
├── agent.yaml          # Agent definition: system prompt, tools, model
├── agent.py            # Agent entrypoint (optional)
└── system_prompt.md    # System prompt template (optional)