config.ts•3.74 kB
import { ServerConfigsSchema } from '../../mcp/config.js';
import { LLMConfigSchema } from '../llm/config.js';
import { EmbeddingConfigSchema } from '../embedding/config.js';
import { z } from 'zod';
export const AgentCardSchema = z
.object({
name: z.string().default('cipher'),
description: z
.string()
.default(
'cipher is an AI assistant capable of store valuable software development knowledge for your vibe coding agents'
),
provider: z
.object({
organization: z.string().default('byterover-inc'),
url: z.string().url().default('https://byterover.dev'),
})
.optional(),
version: z.string().default('1.0.0'),
defaultInputModes: z.array(z.string()).default(['application/json', 'text/plain']),
defaultOutputModes: z
.array(z.string())
.default(['application/json', 'text/event-stream', 'text/plain']),
skills: z
.array(
z.object({
id: z.string(),
name: z.string(),
description: z.string(),
tags: z.array(z.string()),
examples: z.array(z.string()).optional(),
inputModes: z.array(z.string()).optional().default(['text/plain']),
outputModes: z.array(z.string()).optional().default(['text/plain']),
})
)
.default([
{
id: 'chat_with_agent',
name: 'chat_with_agent',
description: 'Allows you to chat with an AI agent. Send a message to interact.',
tags: ['chat', 'AI', 'assistant', 'mcp', 'natural language'],
examples: [
`Send a JSON-RPC request to /mcp with method: "chat_with_agent" and params: {"message":"Your query..."}`,
'Alternatively, use a compatible MCP client library.',
],
},
]),
})
.strict();
export const EventPersistenceConfigSchema = z.object({
enabled: z.boolean().default(true),
storageType: z.enum(['file', 'memory', 'database']).default('file'),
maxEvents: z.number().optional(),
rotationSize: z.number().optional(),
retentionDays: z.number().optional(),
filePath: z.string().optional(),
});
export const AgentConfigSchema = z
.object({
agentCard: AgentCardSchema.describe('Configuration for the agent card').optional(),
systemPrompt: z
.string()
.describe(
'The system prompt content as a string, or a structured system prompt configuration'
),
mcpServers: ServerConfigsSchema.default({}).describe(
'Configurations for MCP (Model Context Protocol) servers used by the agent'
),
llm: LLMConfigSchema.describe('Core LLM configuration for the agent'),
evalLlm: LLMConfigSchema.optional().describe(
'Evaluation LLM configuration for non-thinking tasks (optional, falls back to main LLM if not provided)'
),
embedding: z
.union([EmbeddingConfigSchema, z.object({ disabled: z.boolean() }), z.boolean(), z.null()])
.optional()
.describe(
'Embedding configuration for the agent (optional, falls back to environment auto-detection if not provided). Set to false, null, or {disabled: true} to disable embeddings.'
),
sessions: z
.object({
maxSessions: z
.number()
.int()
.positive()
.default(100)
.describe('Maximum number of concurrent sessions allowed, defaults to 100'),
sessionTTL: z
.number()
.int()
.positive()
.default(3600000)
.describe('Session time-to-live in milliseconds, defaults to 3600000ms (1 hour)'),
})
.default({
maxSessions: 100,
sessionTTL: 3600000,
})
.describe('Session management configuration'),
eventPersistence: EventPersistenceConfigSchema.optional(),
})
.strict()
.describe('Main configuration for an agent, including its LLM and server connections');
// Input type for user-facing API (pre-parsing) - makes fields with defaults optional
export type AgentConfig = z.input<typeof AgentConfigSchema>;