feat: add infinite agent memory with MCP integration

Implement comprehensive conversation management system enabling AI agents
like Claude Code to maintain infinite context and history. Provides semantic
search, smart context retrieval, and automatic artifact linking using Brainy's
existing Triple Intelligence infrastructure.

Core Features:
- ConversationManager API for message storage and retrieval
- MCP protocol integration with 6 tools for Claude Code
- Context ranking using semantic, temporal, and graph scoring
- Neural clustering for theme discovery and deduplication
- Virtual filesystem integration for code artifact linking
- CLI commands for setup and management

Zero new infrastructure required - uses existing Brainy features:
- Storage via brain.add() with NounType.Message
- Relationships via brain.relate() with VerbType.Precedes
- Search via brain.find() with Triple Intelligence
- Clustering via brain.neural()
- Artifacts via brain.vfs()

One-command setup: brainy conversation setup

Version: 3.19.0
This commit is contained in:
David Snelling 2025-09-29 15:37:11 -07:00
parent e3a21c6075
commit ced639cab1
15 changed files with 3304 additions and 7 deletions

View file

@ -0,0 +1,598 @@
/**
* MCP Conversation Tools
*
* Exposes ConversationManager functionality through MCP for Claude Code integration.
* Provides 6 tools for infinite agent memory.
*
* REAL IMPLEMENTATION - Uses ConversationManager which uses real Brainy APIs
*/
import { v4 as uuidv4 } from '../universal/uuid.js'
import {
MCPResponse,
MCPToolExecutionRequest,
MCPTool,
MCP_VERSION
} from '../types/mcpTypes.js'
import { ConversationManager } from '../conversation/conversationManager.js'
import { Brainy } from '../brainy.js'
/**
* MCP Conversation Toolset
*
* Provides conversation and context management tools for AI agents
*/
export class MCPConversationToolset {
private conversationManager: ConversationManager
private initialized = false
/**
* Create MCP Conversation Toolset
* @param brain Brainy instance
*/
constructor(private brain: Brainy) {
this.conversationManager = new ConversationManager(brain)
}
/**
* Initialize the toolset
*/
async init(): Promise<void> {
if (this.initialized) {
return
}
await this.conversationManager.init()
this.initialized = true
}
/**
* Handle MCP tool execution request
* @param request MCP tool execution request
* @returns MCP response
*/
async handleRequest(request: MCPToolExecutionRequest): Promise<MCPResponse> {
if (!this.initialized) {
await this.init()
}
try {
const { toolName, parameters } = request
// Route to appropriate tool handler
switch (toolName) {
case 'conversation_save_message':
return await this.handleSaveMessage(request.requestId, parameters)
case 'conversation_get_context':
return await this.handleGetContext(request.requestId, parameters)
case 'conversation_search':
return await this.handleSearch(request.requestId, parameters)
case 'conversation_get_thread':
return await this.handleGetThread(request.requestId, parameters)
case 'conversation_save_artifact':
return await this.handleSaveArtifact(request.requestId, parameters)
case 'conversation_find_similar':
return await this.handleFindSimilar(request.requestId, parameters)
default:
return this.createErrorResponse(
request.requestId,
'UNKNOWN_TOOL',
`Unknown conversation tool: ${toolName}`
)
}
} catch (error) {
return this.createErrorResponse(
request.requestId,
'INTERNAL_ERROR',
error instanceof Error ? error.message : String(error)
)
}
}
/**
* Get available conversation tools
* @returns Array of MCP tool definitions
*/
async getAvailableTools(): Promise<MCPTool[]> {
return [
{
name: 'conversation_save_message',
description: 'Save a message to conversation history with automatic embedding and indexing',
parameters: {
type: 'object',
properties: {
content: {
type: 'string',
description: 'Message content'
},
role: {
type: 'string',
enum: ['user', 'assistant', 'system', 'tool'],
description: 'Message role'
},
conversationId: {
type: 'string',
description: 'Conversation ID (auto-generated if not provided)'
},
sessionId: {
type: 'string',
description: 'Session ID (optional)'
},
phase: {
type: 'string',
enum: [
'understanding',
'analysis',
'planning',
'implementation',
'testing',
'debugging',
'refinement',
'completed'
],
description: 'Problem-solving phase'
},
confidence: {
type: 'number',
minimum: 0,
maximum: 1,
description: 'Confidence score (0-1)'
},
artifacts: {
type: 'array',
items: { type: 'string' },
description: 'Artifact IDs or paths'
},
toolsUsed: {
type: 'array',
items: { type: 'string' },
description: 'Names of tools used'
},
tags: {
type: 'array',
items: { type: 'string' },
description: 'Tags for categorization'
},
linkToPrevious: {
type: 'string',
description: 'ID of previous message to link'
}
},
required: ['content', 'role']
}
},
{
name: 'conversation_get_context',
description: 'Retrieve relevant context from conversation history using semantic search',
parameters: {
type: 'object',
properties: {
query: {
type: 'string',
description: 'Query string for context retrieval'
},
conversationId: {
type: 'string',
description: 'Limit to specific conversation'
},
limit: {
type: 'number',
description: 'Maximum messages to return (default: 10)',
default: 10
},
maxTokens: {
type: 'number',
description: 'Token budget for context (default: 50000)',
default: 50000
},
relevanceThreshold: {
type: 'number',
minimum: 0,
maximum: 1,
description: 'Minimum similarity score (default: 0.7)',
default: 0.7
},
role: {
oneOf: [
{ type: 'string', enum: ['user', 'assistant', 'system', 'tool'] },
{ type: 'array', items: { type: 'string' } }
],
description: 'Filter by message role'
},
tags: {
type: 'array',
items: { type: 'string' },
description: 'Filter by tags'
},
includeArtifacts: {
type: 'boolean',
description: 'Include linked artifacts',
default: false
},
includeSimilarConversations: {
type: 'boolean',
description: 'Include similar past conversations',
default: false
}
},
required: ['query']
}
},
{
name: 'conversation_search',
description: 'Search messages semantically across all conversations',
parameters: {
type: 'object',
properties: {
query: {
type: 'string',
description: 'Search query'
},
limit: {
type: 'number',
description: 'Maximum results (default: 10)',
default: 10
},
conversationId: {
type: 'string',
description: 'Limit to specific conversation'
},
role: {
oneOf: [
{ type: 'string', enum: ['user', 'assistant', 'system', 'tool'] },
{ type: 'array', items: { type: 'string' } }
],
description: 'Filter by role'
},
timeRange: {
type: 'object',
properties: {
start: { type: 'number', description: 'Start timestamp' },
end: { type: 'number', description: 'End timestamp' }
},
description: 'Time range filter'
}
},
required: ['query']
}
},
{
name: 'conversation_get_thread',
description: 'Get full conversation thread with all messages',
parameters: {
type: 'object',
properties: {
conversationId: {
type: 'string',
description: 'Conversation ID'
},
includeArtifacts: {
type: 'boolean',
description: 'Include linked artifacts',
default: false
}
},
required: ['conversationId']
}
},
{
name: 'conversation_save_artifact',
description: 'Save code/file artifact and link to conversation',
parameters: {
type: 'object',
properties: {
path: {
type: 'string',
description: 'VFS path for artifact'
},
content: {
type: 'string',
description: 'Artifact content'
},
conversationId: {
type: 'string',
description: 'Conversation ID'
},
messageId: {
type: 'string',
description: 'Message ID to link artifact to'
},
type: {
type: 'string',
enum: ['code', 'config', 'data', 'document', 'other'],
description: 'Artifact type'
},
language: {
type: 'string',
description: 'Programming language (for code artifacts)'
},
description: {
type: 'string',
description: 'Artifact description'
}
},
required: ['path', 'content', 'conversationId']
}
},
{
name: 'conversation_find_similar',
description: 'Find similar past conversations using semantic similarity',
parameters: {
type: 'object',
properties: {
conversationId: {
type: 'string',
description: 'Conversation ID to find similar to'
},
limit: {
type: 'number',
description: 'Maximum results (default: 5)',
default: 5
},
threshold: {
type: 'number',
minimum: 0,
maximum: 1,
description: 'Minimum similarity threshold (default: 0.7)',
default: 0.7
}
},
required: ['conversationId']
}
}
]
}
/**
* Handle save_message tool
* REAL: Uses ConversationManager.saveMessage()
*/
private async handleSaveMessage(
requestId: string,
parameters: any
): Promise<MCPResponse> {
const {
content,
role,
conversationId,
sessionId,
phase,
confidence,
artifacts,
toolsUsed,
tags,
linkToPrevious
} = parameters
// Validate required parameters
if (!content || !role) {
return this.createErrorResponse(
requestId,
'INVALID_PARAMETERS',
'Missing required parameters: content and role are required'
)
}
// Save message (REAL)
const messageId = await this.conversationManager.saveMessage(content, role, {
conversationId,
sessionId,
phase,
confidence,
artifacts,
toolsUsed,
tags,
linkToPrevious
})
return this.createSuccessResponse(requestId, {
messageId,
conversationId: conversationId || messageId.split('_')[0]
})
}
/**
* Handle get_context tool
* REAL: Uses ConversationManager.getRelevantContext()
*/
private async handleGetContext(
requestId: string,
parameters: any
): Promise<MCPResponse> {
const { query, ...options } = parameters
if (!query) {
return this.createErrorResponse(
requestId,
'INVALID_PARAMETERS',
'Missing required parameter: query'
)
}
// Get context (REAL)
const context = await this.conversationManager.getRelevantContext(query, options)
return this.createSuccessResponse(requestId, context)
}
/**
* Handle search tool
* REAL: Uses ConversationManager.searchMessages()
*/
private async handleSearch(
requestId: string,
parameters: any
): Promise<MCPResponse> {
const { query } = parameters
if (!query) {
return this.createErrorResponse(
requestId,
'INVALID_PARAMETERS',
'Missing required parameter: query'
)
}
// Search messages (REAL)
const results = await this.conversationManager.searchMessages(parameters)
return this.createSuccessResponse(requestId, {
results,
count: results.length
})
}
/**
* Handle get_thread tool
* REAL: Uses ConversationManager.getConversationThread()
*/
private async handleGetThread(
requestId: string,
parameters: any
): Promise<MCPResponse> {
const { conversationId, includeArtifacts = false } = parameters
if (!conversationId) {
return this.createErrorResponse(
requestId,
'INVALID_PARAMETERS',
'Missing required parameter: conversationId'
)
}
// Get thread (REAL)
const thread = await this.conversationManager.getConversationThread(
conversationId,
{ includeArtifacts }
)
return this.createSuccessResponse(requestId, thread)
}
/**
* Handle save_artifact tool
* REAL: Uses ConversationManager.saveArtifact()
*/
private async handleSaveArtifact(
requestId: string,
parameters: any
): Promise<MCPResponse> {
const {
path,
content,
conversationId,
messageId,
type,
language,
description
} = parameters
if (!path || !content || !conversationId) {
return this.createErrorResponse(
requestId,
'INVALID_PARAMETERS',
'Missing required parameters: path, content, and conversationId are required'
)
}
// Save artifact (REAL)
const artifactId = await this.conversationManager.saveArtifact(path, content, {
conversationId,
messageId,
type,
language,
description
})
return this.createSuccessResponse(requestId, {
artifactId,
path
})
}
/**
* Handle find_similar tool
* REAL: Uses ConversationManager.findSimilarConversations()
*/
private async handleFindSimilar(
requestId: string,
parameters: any
): Promise<MCPResponse> {
const { conversationId, limit = 5, threshold = 0.7 } = parameters
if (!conversationId) {
return this.createErrorResponse(
requestId,
'INVALID_PARAMETERS',
'Missing required parameter: conversationId'
)
}
// Find similar (REAL)
const similar = await this.conversationManager.findSimilarConversations(
conversationId,
limit,
threshold
)
return this.createSuccessResponse(requestId, {
similar,
count: similar.length
})
}
/**
* Create success response
*/
private createSuccessResponse(requestId: string, data: any): MCPResponse {
return {
success: true,
requestId,
version: MCP_VERSION,
data
}
}
/**
* Create error response
*/
private createErrorResponse(
requestId: string,
code: string,
message: string,
details?: any
): MCPResponse {
return {
success: false,
requestId,
version: MCP_VERSION,
error: {
code,
message,
details
}
}
}
/**
* Generate request ID
*/
generateRequestId(): string {
return uuidv4()
}
}
/**
* Create MCP conversation toolset
* @param brain Brainy instance
* @returns MCPConversationToolset instance
*/
export function createConversationToolset(brain: Brainy): MCPConversationToolset {
return new MCPConversationToolset(brain)
}