@polymorphism-tech/morph-spec 2.2.0 → 2.4.0

package/content/.morph/examples/multi-agent/spec.md

@@ -1,479 +1,479 @@
-# Multi-Agent System - Especificação
-
-## 1. Visão Geral
-
-Sistema de agentes de IA cooperativos usando Microsoft Semantic Kernel para resolver tarefas complexas através de decomposição e especialização.
-
-### 1.1 Objetivos
-
-- Criar sistema escalável de agentes cooperativos
-- Implementar orquestração inteligente de tarefas
-- Fornecer extensibilidade para novos agentes e plugins
-- Manter custos controlados com uso eficiente de tokens
-
-### 1.2 Use Cases
-
-| Use Case | Agentes Envolvidos |
-|----------|-------------------|
-| Pesquisa de mercado | Orchestrator → Research → Analysis → Writing |
-| Análise de documentos | Orchestrator → Research → Analysis |
-| Geração de relatórios | Orchestrator → Research → Writing |
-| Q&A com base de conhecimento | Orchestrator → Research |
-
-## 2. Arquitetura
-
-### 2.1 Componentes Principais
-
-```
-┌─────────────────────────────────────────────────────────────────┐
-│                        API Layer                                 │
-│  ┌─────────────────┐  ┌─────────────────┐                       │
-│  │  ChatEndpoints  │  │  AgentEndpoints │                       │
-│  └─────────────────┘  └─────────────────┘                       │
-└─────────────────────────────────────────────────────────────────┘
-                              │
-                              ▼
-┌─────────────────────────────────────────────────────────────────┐
-│                    Orchestration Layer                           │
-│  ┌─────────────────┐  ┌─────────────────┐  ┌─────────────────┐ │
-│  │ TaskDecomposer  │  │  AgentRouter    │  │ ResultAggregator│ │
-│  └─────────────────┘  └─────────────────┘  └─────────────────┘ │
-└─────────────────────────────────────────────────────────────────┘
-                              │
-                              ▼
-┌─────────────────────────────────────────────────────────────────┐
-│                      Agent Layer                                 │
-│  ┌──────────────┐  ┌──────────────┐  ┌──────────────────────┐  │
-│  │ Research     │  │ Analysis     │  │ Writing              │  │
-│  │ Agent        │  │ Agent        │  │ Agent                │  │
-│  └──────────────┘  └──────────────┘  └──────────────────────┘  │
-└─────────────────────────────────────────────────────────────────┘
-                              │
-                              ▼
-┌─────────────────────────────────────────────────────────────────┐
-│                      Plugin Layer                                │
-│  ┌─────────────┐  ┌─────────────┐  ┌─────────────┐             │
-│  │ WebSearch   │  │ Document    │  │ DataAnalysis│             │
-│  │ Plugin      │  │ Plugin      │  │ Plugin      │             │
-│  └─────────────┘  └─────────────┘  └─────────────┘             │
-└─────────────────────────────────────────────────────────────────┘
-                              │
-                              ▼
-┌─────────────────────────────────────────────────────────────────┐
-│                    Infrastructure Layer                          │
-│  ┌─────────────────┐  ┌─────────────────┐  ┌─────────────────┐ │
-│  │  Azure OpenAI   │  │  Vector Store   │  │  Memory Store   │ │
-│  └─────────────────┘  └─────────────────┘  └─────────────────┘ │
-└─────────────────────────────────────────────────────────────────┘
-```
-
-### 2.2 Fluxo de Dados
-
-```
-User Request
-    │
-    ▼
-┌─────────────────────────────────────────┐
-│           Task Decomposition            │
-│  "Compare cloud providers" →            │
-│  [Research AWS, Research Azure,         │
-│   Research GCP, Compare, Write Report]  │
-└─────────────────────────────────────────┘
-    │
-    ▼
-┌─────────────────────────────────────────┐
-│           Agent Execution               │
-│  Research Agent → findings[]            │
-│  Analysis Agent → insights              │
-│  Writing Agent → report                 │
-└─────────────────────────────────────────┘
-    │
-    ▼
-┌─────────────────────────────────────────┐
-│         Result Aggregation              │
-│  Combine outputs, format, validate      │
-└─────────────────────────────────────────┘
-    │
-    ▼
-Final Response
-```
-
-## 3. Requisitos Funcionais
-
-### 3.1 API
-
-| ID | Requisito |
-|----|-----------|
-| API-001 | Endpoint de chat para interação com usuário |
-| API-002 | Endpoint para consultar status de processamento |
-| API-003 | Endpoint para listar agentes disponíveis |
-| API-004 | Streaming de respostas (SSE) |
-| API-005 | Autenticação via API Key |
-
-### 3.2 Orchestrator
-
-| ID | Requisito |
-|----|-----------|
-| ORCH-001 | Análise de intenção do usuário |
-| ORCH-002 | Decomposição de tarefas complexas |
-| ORCH-003 | Roteamento para agentes especialistas |
-| ORCH-004 | Agregação de resultados |
-| ORCH-005 | Retry automático em falhas |
-| ORCH-006 | Timeout por subtarefa |
-
-### 3.3 Agentes
-
-| ID | Requisito |
-|----|-----------|
-| AGT-001 | Research Agent com acesso a web search |
-| AGT-002 | Research Agent com RAG (vector search) |
-| AGT-003 | Analysis Agent para comparações |
-| AGT-004 | Analysis Agent para insights |
-| AGT-005 | Writing Agent para relatórios |
-| AGT-006 | Writing Agent para sumários |
-
-### 3.4 Plugins
-
-| ID | Requisito |
-|----|-----------|
-| PLG-001 | WebSearch via Bing API |
-| PLG-002 | Document retrieval via vector store |
-| PLG-003 | Calculator para análises numéricas |
-| PLG-004 | Date/Time utilities |
-
-### 3.5 Memória
-
-| ID | Requisito |
-|----|-----------|
-| MEM-001 | Conversation memory por sessão |
-| MEM-002 | Semantic memory para contexto relevante |
-| MEM-003 | Persistência opcional em Cosmos DB |
-
-## 4. Requisitos Não-Funcionais
-
-### 4.1 Performance
-
-| ID | Requisito |
-|----|-----------|
-| PERF-001 | Response streaming para UX fluida |
-| PERF-002 | Timeout máximo de 60s por request |
-| PERF-003 | Execução paralela de subtarefas independentes |
-
-### 4.2 Custo
-
-| ID | Requisito |
-|----|-----------|
-| COST-001 | Limite de tokens por request configurável |
-| COST-002 | Uso de modelos menores para subtarefas simples |
-| COST-003 | Cache de resultados de pesquisa |
-| COST-004 | Tracking de uso de tokens |
-
-### 4.3 Segurança
-
-| ID | Requisito |
-|----|-----------|
-| SEC-001 | API Key authentication |
-| SEC-002 | Rate limiting por API key |
-| SEC-003 | Input sanitization |
-| SEC-004 | Output filtering (PII, harmful content) |
-
-## 5. Agentes em Detalhe
-
-### 5.1 Orchestrator Agent
-
-**System Prompt**:
-```
-You are an orchestrator agent responsible for coordinating other specialized agents.
-
-When you receive a user request:
-1. Analyze the intent and complexity
-2. Decompose into subtasks if needed
-3. Delegate to appropriate specialist agents
-4. Aggregate and format the final response
-
-Available agents:
-- research: For information gathering, web search, document retrieval
-- analysis: For data analysis, comparisons, insight generation
-- writing: For text generation, summaries, reports
-
-Respond with a plan in JSON format:
-{
-  "understanding": "Brief summary of user's request",
-  "subtasks": [
-    {"agent": "research", "task": "specific task description"},
-    {"agent": "analysis", "task": "specific task description"},
-    {"agent": "writing", "task": "specific task description"}
-  ]
-}
-```
-
-### 5.2 Research Agent
-
-**System Prompt**:
-```
-You are a research specialist agent. Your capabilities include:
-- Searching the web for current information
-- Retrieving relevant documents from the knowledge base
-- Summarizing and extracting key information
-
-When researching:
-1. Use web_search for current/recent information
-2. Use document_search for internal knowledge
-3. Synthesize findings into structured data
-
-Always cite your sources and indicate confidence level.
-```
-
-**Plugins**:
-- `web_search(query, max_results)`
-- `document_search(query, limit)`
-
-### 5.3 Analysis Agent
-
-**System Prompt**:
-```
-You are an analysis specialist agent. Your capabilities include:
-- Analyzing data and identifying patterns
-- Comparing options and alternatives
-- Generating insights and recommendations
-
-When analyzing:
-1. Structure the data clearly
-2. Apply appropriate analytical frameworks
-3. Provide evidence-based conclusions
-
-Be objective and highlight trade-offs.
-```
-
-**Plugins**:
-- `calculate(expression)`
-- `compare(items, criteria)`
-
-### 5.4 Writing Agent
-
-**System Prompt**:
-```
-You are a writing specialist agent. Your capabilities include:
-- Generating clear, well-structured text
-- Creating summaries and reports
-- Adapting tone and style for the audience
-
-When writing:
-1. Start with a clear structure
-2. Use appropriate formatting
-3. Be concise but comprehensive
-
-Match the requested format (report, summary, email, etc.).
-```
-
-## 6. Plugins em Detalhe
-
-### 6.1 WebSearchPlugin
-
-```csharp
-public class WebSearchPlugin
-{
-    private readonly BingSearchService _searchService;
-
-    [KernelFunction("web_search")]
-    [Description("Search the web for information on a topic")]
-    public async Task<SearchResults> SearchAsync(
-        [Description("The search query")] string query,
-        [Description("Maximum results to return")] int maxResults = 5)
-    {
-        var results = await _searchService.SearchAsync(query, maxResults);
-        return new SearchResults
-        {
-            Query = query,
-            Items = results.Select(r => new SearchResult
-            {
-                Title = r.Title,
-                Snippet = r.Snippet,
-                Url = r.Url
-            }).ToList()
-        };
-    }
-}
-```
-
-### 6.2 DocumentPlugin
-
-```csharp
-public class DocumentPlugin
-{
-    private readonly IVectorStore _vectorStore;
-    private readonly IEmbeddingService _embeddingService;
-
-    [KernelFunction("document_search")]
-    [Description("Search internal documents using semantic similarity")]
-    public async Task<IEnumerable<DocumentResult>> SearchDocumentsAsync(
-        [Description("The search query")] string query,
-        [Description("Maximum documents")] int limit = 3)
-    {
-        var embedding = await _embeddingService.GetEmbeddingAsync(query);
-        var documents = await _vectorStore.SearchAsync(embedding, limit);
-
-        return documents.Select(d => new DocumentResult
-        {
-            Title = d.Title,
-            Content = d.Content,
-            Score = d.Score,
-            Source = d.Source
-        });
-    }
-}
-```
-
-## 7. APIs
-
-### 7.1 Chat Endpoint
-
-```
-POST /api/chat
-Content-Type: application/json
-
-{
-  "message": "Compare AWS, Azure and GCP for a SaaS startup",
-  "sessionId": "optional-session-id",
-  "stream": true
-}
-
-Response (streaming):
-data: {"type": "thinking", "content": "Analyzing request..."}
-data: {"type": "agent", "agent": "research", "status": "started"}
-data: {"type": "agent", "agent": "research", "status": "completed"}
-data: {"type": "content", "content": "Based on my analysis..."}
-data: {"type": "done", "usage": {"tokens": 1234}}
-```
-
-### 7.2 Agents Endpoint
-
-```
-GET /api/agents
-
-Response:
-{
-  "agents": [
-    {
-      "id": "orchestrator",
-      "name": "Orchestrator Agent",
-      "description": "Coordinates other agents",
-      "capabilities": ["task_decomposition", "routing", "aggregation"]
-    },
-    {
-      "id": "research",
-      "name": "Research Agent",
-      "description": "Information gathering specialist",
-      "capabilities": ["web_search", "document_retrieval", "summarization"]
-    }
-  ]
-}
-```
-
-## 8. Modelo de Dados
-
-### 8.1 Conversation
-
-```csharp
-public class Conversation
-{
-    public string Id { get; set; }
-    public string UserId { get; set; }
-    public List<Message> Messages { get; set; }
-    public DateTime CreatedAt { get; set; }
-    public DateTime UpdatedAt { get; set; }
-}
-
-public class Message
-{
-    public string Id { get; set; }
-    public MessageRole Role { get; set; } // User, Assistant, System
-    public string Content { get; set; }
-    public AgentMetadata? AgentMetadata { get; set; }
-    public TokenUsage? Usage { get; set; }
-    public DateTime Timestamp { get; set; }
-}
-
-public class AgentMetadata
-{
-    public string AgentId { get; set; }
-    public List<string> PluginsUsed { get; set; }
-    public Dictionary<string, object> Metadata { get; set; }
-}
-```
-
-### 8.2 Agent Task
-
-```csharp
-public class AgentTask
-{
-    public string Id { get; set; }
-    public string AgentId { get; set; }
-    public string Description { get; set; }
-    public TaskStatus Status { get; set; }
-    public string? Input { get; set; }
-    public string? Output { get; set; }
-    public string? Error { get; set; }
-    public TimeSpan Duration { get; set; }
-    public TokenUsage Usage { get; set; }
-}
-
-public enum TaskStatus
-{
-    Pending,
-    Running,
-    Completed,
-    Failed
-}
-```
-
-## 9. Configuração
-
-### 9.1 Azure OpenAI
-
-```json
-{
-  "AzureOpenAI": {
-    "Endpoint": "https://{resource}.openai.azure.com/",
-    "ApiKey": "{key}",
-    "Deployments": {
-      "Chat": "gpt-4",
-      "ChatFast": "gpt-35-turbo",
-      "Embedding": "text-embedding-ada-002"
-    }
-  }
-}
-```
-
-### 9.2 Agent Configuration
-
-```json
-{
-  "Agents": {
-    "Orchestrator": {
-      "Model": "gpt-4",
-      "MaxTokens": 2000,
-      "Temperature": 0.3
-    },
-    "Research": {
-      "Model": "gpt-35-turbo",
-      "MaxTokens": 1500,
-      "Temperature": 0.5
-    },
-    "Analysis": {
-      "Model": "gpt-35-turbo",
-      "MaxTokens": 1500,
-      "Temperature": 0.3
-    },
-    "Writing": {
-      "Model": "gpt-4",
-      "MaxTokens": 3000,
-      "Temperature": 0.7
-    }
-  }
-}
-```
-
----
-
-*MORPH-SPEC by Polymorphism Tech*
+# Multi-Agent System - Especificação
+
+## 1. Visão Geral
+
+Sistema de agentes de IA cooperativos usando Microsoft Semantic Kernel para resolver tarefas complexas através de decomposição e especialização.
+
+### 1.1 Objetivos
+
+- Criar sistema escalável de agentes cooperativos
+- Implementar orquestração inteligente de tarefas
+- Fornecer extensibilidade para novos agentes e plugins
+- Manter custos controlados com uso eficiente de tokens
+
+### 1.2 Use Cases
+
+| Use Case | Agentes Envolvidos |
+|----------|-------------------|
+| Pesquisa de mercado | Orchestrator → Research → Analysis → Writing |
+| Análise de documentos | Orchestrator → Research → Analysis |
+| Geração de relatórios | Orchestrator → Research → Writing |
+| Q&A com base de conhecimento | Orchestrator → Research |
+
+## 2. Arquitetura
+
+### 2.1 Componentes Principais
+
+```
+┌─────────────────────────────────────────────────────────────────┐
+│                        API Layer                                 │
+│  ┌─────────────────┐  ┌─────────────────┐                       │
+│  │  ChatEndpoints  │  │  AgentEndpoints │                       │
+│  └─────────────────┘  └─────────────────┘                       │
+└─────────────────────────────────────────────────────────────────┘
+                              │
+                              ▼
+┌─────────────────────────────────────────────────────────────────┐
+│                    Orchestration Layer                           │
+│  ┌─────────────────┐  ┌─────────────────┐  ┌─────────────────┐ │
+│  │ TaskDecomposer  │  │  AgentRouter    │  │ ResultAggregator│ │
+│  └─────────────────┘  └─────────────────┘  └─────────────────┘ │
+└─────────────────────────────────────────────────────────────────┘
+                              │
+                              ▼
+┌─────────────────────────────────────────────────────────────────┐
+│                      Agent Layer                                 │
+│  ┌──────────────┐  ┌──────────────┐  ┌──────────────────────┐  │
+│  │ Research     │  │ Analysis     │  │ Writing              │  │
+│  │ Agent        │  │ Agent        │  │ Agent                │  │
+│  └──────────────┘  └──────────────┘  └──────────────────────┘  │
+└─────────────────────────────────────────────────────────────────┘
+                              │
+                              ▼
+┌─────────────────────────────────────────────────────────────────┐
+│                      Plugin Layer                                │
+│  ┌─────────────┐  ┌─────────────┐  ┌─────────────┐             │
+│  │ WebSearch   │  │ Document    │  │ DataAnalysis│             │
+│  │ Plugin      │  │ Plugin      │  │ Plugin      │             │
+│  └─────────────┘  └─────────────┘  └─────────────┘             │
+└─────────────────────────────────────────────────────────────────┘
+                              │
+                              ▼
+┌─────────────────────────────────────────────────────────────────┐
+│                    Infrastructure Layer                          │
+│  ┌─────────────────┐  ┌─────────────────┐  ┌─────────────────┐ │
+│  │  Azure OpenAI   │  │  Vector Store   │  │  Memory Store   │ │
+│  └─────────────────┘  └─────────────────┘  └─────────────────┘ │
+└─────────────────────────────────────────────────────────────────┘
+```
+
+### 2.2 Fluxo de Dados
+
+```
+User Request
+    │
+    ▼
+┌─────────────────────────────────────────┐
+│           Task Decomposition            │
+│  "Compare cloud providers" →            │
+│  [Research AWS, Research Azure,         │
+│   Research GCP, Compare, Write Report]  │
+└─────────────────────────────────────────┘
+    │
+    ▼
+┌─────────────────────────────────────────┐
+│           Agent Execution               │
+│  Research Agent → findings[]            │
+│  Analysis Agent → insights              │
+│  Writing Agent → report                 │
+└─────────────────────────────────────────┘
+    │
+    ▼
+┌─────────────────────────────────────────┐
+│         Result Aggregation              │
+│  Combine outputs, format, validate      │
+└─────────────────────────────────────────┘
+    │
+    ▼
+Final Response
+```
+
+## 3. Requisitos Funcionais
+
+### 3.1 API
+
+| ID | Requisito |
+|----|-----------|
+| API-001 | Endpoint de chat para interação com usuário |
+| API-002 | Endpoint para consultar status de processamento |
+| API-003 | Endpoint para listar agentes disponíveis |
+| API-004 | Streaming de respostas (SSE) |
+| API-005 | Autenticação via API Key |
+
+### 3.2 Orchestrator
+
+| ID | Requisito |
+|----|-----------|
+| ORCH-001 | Análise de intenção do usuário |
+| ORCH-002 | Decomposição de tarefas complexas |
+| ORCH-003 | Roteamento para agentes especialistas |
+| ORCH-004 | Agregação de resultados |
+| ORCH-005 | Retry automático em falhas |
+| ORCH-006 | Timeout por subtarefa |
+
+### 3.3 Agentes
+
+| ID | Requisito |
+|----|-----------|
+| AGT-001 | Research Agent com acesso a web search |
+| AGT-002 | Research Agent com RAG (vector search) |
+| AGT-003 | Analysis Agent para comparações |
+| AGT-004 | Analysis Agent para insights |
+| AGT-005 | Writing Agent para relatórios |
+| AGT-006 | Writing Agent para sumários |
+
+### 3.4 Plugins
+
+| ID | Requisito |
+|----|-----------|
+| PLG-001 | WebSearch via Bing API |
+| PLG-002 | Document retrieval via vector store |
+| PLG-003 | Calculator para análises numéricas |
+| PLG-004 | Date/Time utilities |
+
+### 3.5 Memória
+
+| ID | Requisito |
+|----|-----------|
+| MEM-001 | Conversation memory por sessão |
+| MEM-002 | Semantic memory para contexto relevante |
+| MEM-003 | Persistência opcional em Cosmos DB |
+
+## 4. Requisitos Não-Funcionais
+
+### 4.1 Performance
+
+| ID | Requisito |
+|----|-----------|
+| PERF-001 | Response streaming para UX fluida |
+| PERF-002 | Timeout máximo de 60s por request |
+| PERF-003 | Execução paralela de subtarefas independentes |
+
+### 4.2 Custo
+
+| ID | Requisito |
+|----|-----------|
+| COST-001 | Limite de tokens por request configurável |
+| COST-002 | Uso de modelos menores para subtarefas simples |
+| COST-003 | Cache de resultados de pesquisa |
+| COST-004 | Tracking de uso de tokens |
+
+### 4.3 Segurança
+
+| ID | Requisito |
+|----|-----------|
+| SEC-001 | API Key authentication |
+| SEC-002 | Rate limiting por API key |
+| SEC-003 | Input sanitization |
+| SEC-004 | Output filtering (PII, harmful content) |
+
+## 5. Agentes em Detalhe
+
+### 5.1 Orchestrator Agent
+
+**System Prompt**:
+```
+You are an orchestrator agent responsible for coordinating other specialized agents.
+
+When you receive a user request:
+1. Analyze the intent and complexity
+2. Decompose into subtasks if needed
+3. Delegate to appropriate specialist agents
+4. Aggregate and format the final response
+
+Available agents:
+- research: For information gathering, web search, document retrieval
+- analysis: For data analysis, comparisons, insight generation
+- writing: For text generation, summaries, reports
+
+Respond with a plan in JSON format:
+{
+  "understanding": "Brief summary of user's request",
+  "subtasks": [
+    {"agent": "research", "task": "specific task description"},
+    {"agent": "analysis", "task": "specific task description"},
+    {"agent": "writing", "task": "specific task description"}
+  ]
+}
+```
+
+### 5.2 Research Agent
+
+**System Prompt**:
+```
+You are a research specialist agent. Your capabilities include:
+- Searching the web for current information
+- Retrieving relevant documents from the knowledge base
+- Summarizing and extracting key information
+
+When researching:
+1. Use web_search for current/recent information
+2. Use document_search for internal knowledge
+3. Synthesize findings into structured data
+
+Always cite your sources and indicate confidence level.
+```
+
+**Plugins**:
+- `web_search(query, max_results)`
+- `document_search(query, limit)`
+
+### 5.3 Analysis Agent
+
+**System Prompt**:
+```
+You are an analysis specialist agent. Your capabilities include:
+- Analyzing data and identifying patterns
+- Comparing options and alternatives
+- Generating insights and recommendations
+
+When analyzing:
+1. Structure the data clearly
+2. Apply appropriate analytical frameworks
+3. Provide evidence-based conclusions
+
+Be objective and highlight trade-offs.
+```
+
+**Plugins**:
+- `calculate(expression)`
+- `compare(items, criteria)`
+
+### 5.4 Writing Agent
+
+**System Prompt**:
+```
+You are a writing specialist agent. Your capabilities include:
+- Generating clear, well-structured text
+- Creating summaries and reports
+- Adapting tone and style for the audience
+
+When writing:
+1. Start with a clear structure
+2. Use appropriate formatting
+3. Be concise but comprehensive
+
+Match the requested format (report, summary, email, etc.).
+```
+
+## 6. Plugins em Detalhe
+
+### 6.1 WebSearchPlugin
+
+```csharp
+public class WebSearchPlugin
+{
+    private readonly BingSearchService _searchService;
+
+    [KernelFunction("web_search")]
+    [Description("Search the web for information on a topic")]
+    public async Task<SearchResults> SearchAsync(
+        [Description("The search query")] string query,
+        [Description("Maximum results to return")] int maxResults = 5)
+    {
+        var results = await _searchService.SearchAsync(query, maxResults);
+        return new SearchResults
+        {
+            Query = query,
+            Items = results.Select(r => new SearchResult
+            {
+                Title = r.Title,
+                Snippet = r.Snippet,
+                Url = r.Url
+            }).ToList()
+        };
+    }
+}
+```
+
+### 6.2 DocumentPlugin
+
+```csharp
+public class DocumentPlugin
+{
+    private readonly IVectorStore _vectorStore;
+    private readonly IEmbeddingService _embeddingService;
+
+    [KernelFunction("document_search")]
+    [Description("Search internal documents using semantic similarity")]
+    public async Task<IEnumerable<DocumentResult>> SearchDocumentsAsync(
+        [Description("The search query")] string query,
+        [Description("Maximum documents")] int limit = 3)
+    {
+        var embedding = await _embeddingService.GetEmbeddingAsync(query);
+        var documents = await _vectorStore.SearchAsync(embedding, limit);
+
+        return documents.Select(d => new DocumentResult
+        {
+            Title = d.Title,
+            Content = d.Content,
+            Score = d.Score,
+            Source = d.Source
+        });
+    }
+}
+```
+
+## 7. APIs
+
+### 7.1 Chat Endpoint
+
+```
+POST /api/chat
+Content-Type: application/json
+
+{
+  "message": "Compare AWS, Azure and GCP for a SaaS startup",
+  "sessionId": "optional-session-id",
+  "stream": true
+}
+
+Response (streaming):
+data: {"type": "thinking", "content": "Analyzing request..."}
+data: {"type": "agent", "agent": "research", "status": "started"}
+data: {"type": "agent", "agent": "research", "status": "completed"}
+data: {"type": "content", "content": "Based on my analysis..."}
+data: {"type": "done", "usage": {"tokens": 1234}}
+```
+
+### 7.2 Agents Endpoint
+
+```
+GET /api/agents
+
+Response:
+{
+  "agents": [
+    {
+      "id": "orchestrator",
+      "name": "Orchestrator Agent",
+      "description": "Coordinates other agents",
+      "capabilities": ["task_decomposition", "routing", "aggregation"]
+    },
+    {
+      "id": "research",
+      "name": "Research Agent",
+      "description": "Information gathering specialist",
+      "capabilities": ["web_search", "document_retrieval", "summarization"]
+    }
+  ]
+}
+```
+
+## 8. Modelo de Dados
+
+### 8.1 Conversation
+
+```csharp
+public class Conversation
+{
+    public string Id { get; set; }
+    public string UserId { get; set; }
+    public List<Message> Messages { get; set; }
+    public DateTime CreatedAt { get; set; }
+    public DateTime UpdatedAt { get; set; }
+}
+
+public class Message
+{
+    public string Id { get; set; }
+    public MessageRole Role { get; set; } // User, Assistant, System
+    public string Content { get; set; }
+    public AgentMetadata? AgentMetadata { get; set; }
+    public TokenUsage? Usage { get; set; }
+    public DateTime Timestamp { get; set; }
+}
+
+public class AgentMetadata
+{
+    public string AgentId { get; set; }
+    public List<string> PluginsUsed { get; set; }
+    public Dictionary<string, object> Metadata { get; set; }
+}
+```
+
+### 8.2 Agent Task
+
+```csharp
+public class AgentTask
+{
+    public string Id { get; set; }
+    public string AgentId { get; set; }
+    public string Description { get; set; }
+    public TaskStatus Status { get; set; }
+    public string? Input { get; set; }
+    public string? Output { get; set; }
+    public string? Error { get; set; }
+    public TimeSpan Duration { get; set; }
+    public TokenUsage Usage { get; set; }
+}
+
+public enum TaskStatus
+{
+    Pending,
+    Running,
+    Completed,
+    Failed
+}
+```
+
+## 9. Configuração
+
+### 9.1 Azure OpenAI
+
+```json
+{
+  "AzureOpenAI": {
+    "Endpoint": "https://{resource}.openai.azure.com/",
+    "ApiKey": "{key}",
+    "Deployments": {
+      "Chat": "gpt-4",
+      "ChatFast": "gpt-35-turbo",
+      "Embedding": "text-embedding-ada-002"
+    }
+  }
+}
+```
+
+### 9.2 Agent Configuration
+
+```json
+{
+  "Agents": {
+    "Orchestrator": {
+      "Model": "gpt-4",
+      "MaxTokens": 2000,
+      "Temperature": 0.3
+    },
+    "Research": {
+      "Model": "gpt-35-turbo",
+      "MaxTokens": 1500,
+      "Temperature": 0.5
+    },
+    "Analysis": {
+      "Model": "gpt-35-turbo",
+      "MaxTokens": 1500,
+      "Temperature": 0.3
+    },
+    "Writing": {
+      "Model": "gpt-4",
+      "MaxTokens": 3000,
+      "Temperature": 0.7
+    }
+  }
+}
+```
+
+---
+
+*MORPH-SPEC by Polymorphism Tech*