Skip to main content

Type Alias: WunderlandAgentConfig

WunderlandAgentConfig = object

Defined in: .source/wunderland/src/channels/api/types.ts:346

Minimal shape of the agent.config.json schema used by Wunderland CLI/runtime. This is the "control-plane" config you export from dashboards and run with wunderland start.

Properties​

adaptiveExecution?​

optional adaptiveExecution?: WunderlandAdaptiveExecutionConfig

Defined in: .source/wunderland/src/channels/api/types.ts:471

Adaptive execution controls based on rolling KPI.

agentName?​

optional agentName?: string

Defined in: .source/wunderland/src/channels/api/types.ts:351

Alias for displayName — used by global CLI config (~/.wunderland/config.json).

bio?​

optional bio?: string

Defined in: .source/wunderland/src/channels/api/types.ts:352

channels?​

optional channels?: string[]

Defined in: .source/wunderland/src/channels/api/types.ts:403

Channel platform IDs (e.g., "telegram").

discovery?​

optional discovery?: WunderlandAgentDiscoveryConfig

Defined in: .source/wunderland/src/channels/api/types.ts:477

Capability discovery configuration.

displayName?​

optional displayName?: string

Defined in: .source/wunderland/src/channels/api/types.ts:349

executionMode?​

optional executionMode?: WunderlandExecutionMode | string | null

Defined in: .source/wunderland/src/channels/api/types.ts:390

extensionOverrides?​

optional extensionOverrides?: Record<string, { enabled?: boolean; options?: unknown; priority?: number; }>

Defined in: .source/wunderland/src/channels/api/types.ts:394

extensions?​

optional extensions?: WunderlandExtensionConfig

Defined in: .source/wunderland/src/channels/api/types.ts:393

hitl?​

optional hitl?: Partial<{ secret: string; turnApproval: "off" | "after-each-turn" | "after-each-round" | string; turnApprovalMode: "off" | "after-each-turn" | "after-each-round" | string; }>

Defined in: .source/wunderland/src/channels/api/types.ts:441

lazyTools?​

optional lazyTools?: boolean

Defined in: .source/wunderland/src/channels/api/types.ts:391

llmAuthMethod?​

optional llmAuthMethod?: "api-key" | "oauth"

Defined in: .source/wunderland/src/channels/api/types.ts:367

Auth method for the LLM provider. Use 'api-key'. 'oauth' remains for legacy config detection and is rejected at runtime.

llmModel?​

optional llmModel?: string

Defined in: .source/wunderland/src/channels/api/types.ts:365

llmProvider?​

optional llmProvider?: string

Defined in: .source/wunderland/src/channels/api/types.ts:364

memory?​

optional memory?: object

Defined in: .source/wunderland/src/channels/api/types.ts:408

Memory retrieval configuration.

cognitiveMechanisms?​

optional cognitiveMechanisms?: any

Optional cognitive mechanisms (reconsolidation, RIF, FOK, etc.). When present, CognitiveMemoryManager is initialized and auto-ingest facts route through cognitive encoding.

Default​
undefined (disabled)

enabled?​

optional enabled?: boolean

Enable/disable memory retrieval.

Default​
true

infiniteContext?​

optional infiniteContext?: object

Infinite context window config.

infiniteContext.compactionThreshold?​

optional compactionThreshold?: number

infiniteContext.enabled?​

optional enabled?: boolean

infiniteContext.preserveRecentTurns?​

optional preserveRecentTurns?: number

infiniteContext.strategy?​

optional strategy?: "sliding" | "hierarchical" | "hybrid"

infiniteContext.transparencyLevel?​

optional transparencyLevel?: "none" | "summary" | "detailed"

maxContextTokens?​

optional maxContextTokens?: number

Max context tokens for context window manager.

retrievalBudgetTokens?​

optional retrievalBudgetTokens?: number

Total token budget for memory context per turn.

Default​
4000

observability?​

optional observability?: Partial<{ otel?: Partial<{ enabled: boolean; exportLogs: boolean; }>; textLogs?: Partial<{ directory: string; enabled: boolean; includeToolCalls: boolean; }>; }>

Defined in: .source/wunderland/src/channels/api/types.ts:453

ollama?​

optional ollama?: Partial<{ baseUrl: string; numCtx: number; numGpu: number; }>

Defined in: .source/wunderland/src/channels/api/types.ts:369

Ollama-specific runtime settings.

pairing?​

optional pairing?: Partial<{ codeLength: number; enabled: boolean; groupTrigger: string; maxPending: number; pendingTtlMs: number; }>

Defined in: .source/wunderland/src/channels/api/types.ts:446

permissionSet?​

optional permissionSet?: PermissionSetName | string | null

Defined in: .source/wunderland/src/channels/api/types.ts:388

personality?​

optional personality?: Partial<{ agreeableness: number; conscientiousness: number; emotionality: number; extraversion: number; honesty: number; openness: number; }>

Defined in: .source/wunderland/src/channels/api/types.ts:356

personaRegistry?​

optional personaRegistry?: WunderlandAgentPersonaRegistryConfig

Defined in: .source/wunderland/src/channels/api/types.ts:501

AgentOS persona registry configuration.

presetId?​

optional presetId?: string

Defined in: .source/wunderland/src/channels/api/types.ts:348

providerPreferences?​

optional providerPreferences?: object

Defined in: .source/wunderland/src/channels/api/types.ts:503

User-configurable provider ordering for media generation.

audio?​

optional audio?: object

audio.music?​

optional music?: object

audio.music.blocked?​

optional blocked?: string[]

audio.music.preferred?​

optional preferred?: string[]

audio.music.weights?​

optional weights?: Record<string, number>

audio.sfx?​

optional sfx?: object

audio.sfx.blocked?​

optional blocked?: string[]

audio.sfx.preferred?​

optional preferred?: string[]

audio.sfx.weights?​

optional weights?: Record<string, number>

image?​

optional image?: object

image.blocked?​

optional blocked?: string[]

image.preferred?​

optional preferred?: string[]

image.weights?​

optional weights?: Record<string, number>

video?​

optional video?: object

video.blocked?​

optional blocked?: string[]

video.preferred?​

optional preferred?: string[]

video.weights?​

optional weights?: Record<string, number>

rag?​

optional rag?: WunderlandAgentRagConfig

Defined in: .source/wunderland/src/channels/api/types.ts:497

RAG / long-term memory configuration.

research?​

optional research?: WunderlandAgentResearchConfig

Defined in: .source/wunderland/src/channels/api/types.ts:499

Research depth escalation / classifier configuration.

secrets?​

optional secrets?: Record<string, string>

Defined in: .source/wunderland/src/channels/api/types.ts:406

Optional secret overrides (in addition to env vars).

security?​

optional security?: Partial<{ dualLLMAudit: boolean; dualLlmAuditor: boolean; folderPermissions: FolderPermissionConfig; outputSigning: boolean; permissionSet: PermissionSetName | string; preLlmClassifier: boolean; preLLMClassifier: boolean; storagePolicy: string; tier: SecurityTierName | string; wrapToolOutputs: boolean; }>

Defined in: .source/wunderland/src/channels/api/types.ts:429

Optional security overrides that the runtime reads.

securityTier?​

optional securityTier?: SecurityTierName | string | null

Defined in: .source/wunderland/src/channels/api/types.ts:387

seedId?​

optional seedId?: string

Defined in: .source/wunderland/src/channels/api/types.ts:347

selectedPersonaId?​

optional selectedPersonaId?: string

Defined in: .source/wunderland/src/channels/api/types.ts:355

Default selected AgentOS persona ID applied at runtime/tool context scope.

skills?​

optional skills?: string[]

Defined in: .source/wunderland/src/channels/api/types.ts:392

storage?​

optional storage?: object

Defined in: .source/wunderland/src/channels/api/types.ts:479

Per-agent storage configuration (SQLite or cloud).

autoIngest?​

optional autoIngest?: object

Auto-ingest pipeline configuration.

autoIngest.enabled?​

optional enabled?: boolean

Enable automatic memory extraction from conversations (default: true).

autoIngest.importanceThreshold?​

optional importanceThreshold?: number

Override personality-derived importance threshold (0.0-1.0).

autoIngest.maxPerTurn?​

optional maxPerTurn?: number

Max memories to extract per conversation turn (default: 3).

backend?​

optional backend?: "local" | "cloud"

Storage backend: 'local' (SQLite per agent, default) or 'cloud' (Postgres/Supabase).

connectionString?​

optional connectionString?: string

Connection string for cloud backend (reads DATABASE_URL env if not set).

dbPath?​

optional dbPath?: string

Override default DB path (~/.wunderland/agents/{seedId}/agent.db).

suggestedChannels?​

optional suggestedChannels?: string[]

Defined in: .source/wunderland/src/channels/api/types.ts:404

systemPrompt?​

optional systemPrompt?: string

Defined in: .source/wunderland/src/channels/api/types.ts:353

taskOutcomeTelemetry?​

optional taskOutcomeTelemetry?: WunderlandTaskOutcomeTelemetryConfig

Defined in: .source/wunderland/src/channels/api/types.ts:467

Rolling task-outcome KPI telemetry.

toolAccessProfile?​

optional toolAccessProfile?: ToolAccessProfileName | string | null

Defined in: .source/wunderland/src/channels/api/types.ts:389

toolCalling?​

optional toolCalling?: Partial<{ strictToolNames: boolean; }>

Defined in: .source/wunderland/src/channels/api/types.ts:380

toolFailureMode?​

optional toolFailureMode?: WunderlandToolFailureMode | string | null

Defined in: .source/wunderland/src/channels/api/types.ts:379

Tool-call failure behavior:

  • fail_open: continue after tool failures (default)
  • fail_closed: stop turn on first tool failure

wallet?​

optional wallet?: WalletConfig

Defined in: .source/wunderland/src/channels/api/types.ts:474

Agent personal wallet configuration (crypto + virtual cards).