agents

Classes

AgenticAdapter

Defined in: packages/nexus-agents/src/agents/agentic/agentic-adapter.ts:97

Agents exports - Agent framework, Orchestrator, Experts Split from index.ts for file size compliance (Issue #285)

Implements

Constructors

Constructor
new AgenticAdapter(modelAdapter, options?): AgenticAdapter;

Defined in: packages/nexus-agents/src/agents/agentic/agentic-adapter.ts:120

Parameters
modelAdapter

IModelAdapter

options?

AgenticAdapterOptions = {}

Returns

AgenticAdapter

Properties

adapterStrategy
adapterStrategy: string;

Defined in: packages/nexus-agents/src/agents/agentic/agentic-adapter.ts:110

Adapter strategy stamp — composed from the resolved model identity, NOT the IModelAdapter’s providerId. For a custom OpenAI gateway fronting Claude, this reads native:anthropic even though IModelAdapter.providerId === 'openai'.

Initialised eagerly from the modelId parse (sync); upgraded after the first runAgent if the probe contributes a higher-confidence vendor signal.

Implementation of

IAgenticAdapter.adapterStrategy

modelId
readonly modelId: string;

Defined in: packages/nexus-agents/src/agents/agentic/agentic-adapter.ts:99

Implementation of

IAgenticAdapter.modelId

providerId
readonly providerId: string;

Defined in: packages/nexus-agents/src/agents/agentic/agentic-adapter.ts:98

Implementation of

IAgenticAdapter.providerId

Methods

getProfile()
getProfile(): ModelEntry;

Defined in: packages/nexus-agents/src/agents/agentic/agentic-adapter.ts:152

Read-only accessor for the resolved profile. Mostly used in tests

  • observability surfaces; production callers shouldn’t need this.
Returns

ModelEntry

getResolvedIdentity()
getResolvedIdentity(): ResolvedModelIdentity;

Defined in: packages/nexus-agents/src/agents/agentic/agentic-adapter.ts:161

Read-only accessor for the resolved identity. After the first runAgent call (if a probe ran), this may differ from what the sync constructor stored — useful for audit logs.

Returns

ResolvedModelIdentity

runAgent()
runAgent(args): Promise<Result<AgentRunResult, AgentError>>;

Defined in: packages/nexus-agents/src/agents/agentic/agentic-adapter.ts:228

Parameters
args

RunAgentArgs

Returns

Promise<Result<AgentRunResult, AgentError>>

Implementation of

IAgenticAdapter.runAgent


AgentStateMachine

Defined in: packages/nexus-agents/src/agents/state-machine.ts:42

Agent State Machine.

Manages agent lifecycle states with validation, event callbacks, and error recovery mechanisms.

Constructors

Constructor
new AgentStateMachine(options?): AgentStateMachine;

Defined in: packages/nexus-agents/src/agents/state-machine.ts:53

Parameters
options?

StateMachineOptions = {}

Returns

AgentStateMachine

Accessors

errors
Get Signature
get errors(): number;

Defined in: packages/nexus-agents/src/agents/state-machine.ts:77

Gets the current error count.

Returns

number

state
Get Signature
get state(): AgentState;

Defined in: packages/nexus-agents/src/agents/state-machine.ts:63

Gets the current state.

Returns

AgentState

transitionHistory
Get Signature
get transitionHistory(): readonly StateTransition[];

Defined in: packages/nexus-agents/src/agents/state-machine.ts:70

Gets the transition history.

Returns

readonly StateTransition[]

Methods

canTransition()
canTransition(event): boolean;

Defined in: packages/nexus-agents/src/agents/state-machine.ts:84

Checks if a transition is valid from the current state.

Parameters
event

StateTransitionEvent

Returns

boolean

forceError()
forceError(context?): void;

Defined in: packages/nexus-agents/src/agents/state-machine.ts:146

Forces a transition to the error state. Use for unrecoverable errors that should bypass normal transition rules.

Parameters
context?

Record<string, unknown>

Optional context data about the error

Returns

void

getNextState()
getNextState(event): AgentState | undefined;

Defined in: packages/nexus-agents/src/agents/state-machine.ts:92

Gets the next state for an event, if valid.

Parameters
event

StateTransitionEvent

Returns

AgentState | undefined

getValidEvents()
getValidEvents(): StateTransitionEvent[];

Defined in: packages/nexus-agents/src/agents/state-machine.ts:100

Gets all valid events from the current state.

Returns

StateTransitionEvent[]

hasError()
hasError(): boolean;

Defined in: packages/nexus-agents/src/agents/state-machine.ts:250

Checks if the agent is in an error state.

Returns

boolean

isAvailable()
isAvailable(): boolean;

Defined in: packages/nexus-agents/src/agents/state-machine.ts:236

Checks if the agent is in a state where it can accept new tasks.

Returns

boolean

isWorking()
isWorking(): boolean;

Defined in: packages/nexus-agents/src/agents/state-machine.ts:243

Checks if the agent is currently working.

Returns

boolean

onStateChange()
onStateChange(callback): () => void;

Defined in: packages/nexus-agents/src/agents/state-machine.ts:213

Subscribes to state change events.

Parameters
callback

StateChangeCallback

Callback to invoke on state changes

Returns

Unsubscribe function

() => void

onTransitionError()
onTransitionError(callback): () => void;

Defined in: packages/nexus-agents/src/agents/state-machine.ts:226

Subscribes to transition error events.

Parameters
callback

TransitionErrorCallback

Callback to invoke on transition errors

Returns

Unsubscribe function

() => void

recover()
recover(context?): Result<AgentState, AgentError>;

Defined in: packages/nexus-agents/src/agents/state-machine.ts:161

Attempts recovery from the error state.

Parameters
context?

Record<string, unknown>

Optional context data about the recovery

Returns

Result<AgentState, AgentError>

Result with the new state or an AgentError if recovery failed

reset()
reset(clearHistory?): void;

Defined in: packages/nexus-agents/src/agents/state-machine.ts:199

Resets the state machine to its initial state.

Parameters
clearHistory?

boolean = false

Whether to clear the transition history

Returns

void

resetErrorCount()
resetErrorCount(): void;

Defined in: packages/nexus-agents/src/agents/state-machine.ts:190

Resets the error count. Call after successful task completion.

Returns

void

transition()
transition(event, context?): Result<AgentState, AgentError>;

Defined in: packages/nexus-agents/src/agents/state-machine.ts:112

Attempts a state transition.

Parameters
event

StateTransitionEvent

The event triggering the transition

context?

Record<string, unknown>

Optional context data for the transition

Returns

Result<AgentState, AgentError>

Result with the new state or an AgentError


ArchitectureExpert

Defined in: packages/nexus-agents/src/agents/experts/architecture-expert.ts:64

ArchitectureExpert - Expert agent for architecture-related tasks.

Extends

Constructors

Constructor
new ArchitectureExpert(options?): ArchitectureExpert;

Defined in: packages/nexus-agents/src/agents/experts/architecture-expert.ts:67

Parameters
options?

Partial<BaseAgentOptions> & { expertOptions?: ArchitectureExpertOptions; } = {}

Returns

ArchitectureExpert

Overrides

BaseAgent.constructor

Properties

adapter
protected adapter: IModelAdapter | undefined;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:98

Inherited from

BaseAgent.adapter

budgetTracker
protected readonly budgetTracker: ITokenBudgetTracker;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:97

Inherited from

BaseAgent.budgetTracker

capabilities
readonly capabilities: readonly AgentCapability[];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:95

Agent capabilities

Inherited from

BaseAgent.capabilities

config
protected config: AgentConfig | undefined;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:100

Inherited from

BaseAgent.config

emitMessageEvents
protected readonly emitMessageEvents: boolean;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:107

Inherited from

BaseAgent.emitMessageEvents

eventBus
protected readonly eventBus: ICollaborationEventBus;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:106

Inherited from

BaseAgent.eventBus

history
protected history: Message[] = [];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:102

Inherited from

BaseAgent.history

id
readonly id: string;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:93

Unique agent identifier

Inherited from

BaseAgent.id

logger
protected readonly logger: ILogger;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:99

Inherited from

BaseAgent.logger

maxTokens
protected readonly maxTokens: number;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:105

Inherited from

BaseAgent.maxTokens

role
readonly role: AgentRole;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:94

Agent role

Inherited from

BaseAgent.role

sharedState
protected sharedState: Record<string, unknown> = {};

Defined in: packages/nexus-agents/src/agents/base-agent.ts:101

Inherited from

BaseAgent.sharedState

stateMachine
protected readonly stateMachine: AgentStateMachine;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:96

Inherited from

BaseAgent.stateMachine

systemPrompt
protected readonly systemPrompt: string | undefined;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:103

Inherited from

BaseAgent.systemPrompt

temperature
protected readonly temperature: number;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:104

Inherited from

BaseAgent.temperature

Accessors

state
Get Signature
get state(): AgentState;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:176

Current state

Returns

AgentState

Current state

Inherited from

BaseAgent.state

Methods

addContextItem()
protected addContextItem(
   content, 
   priority?, 
   category?
): Promise<void>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:344

Parameters
content

string

priority?

100 | 20 | 60 | 80 | 40

category?

"system" | "task" | "active"

Returns

Promise<void>

Inherited from

BaseAgent.addContextItem

addToHistory()
protected addToHistory(message): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:331

Parameters
message

Message

Returns

void

Inherited from

BaseAgent.addToHistory

buildPrompt()
protected buildPrompt(task): Message[];

Defined in: packages/nexus-agents/src/agents/experts/architecture-expert.ts:95

Parameters
task

Task

Returns

Message[]

Overrides

BaseAgent.buildPrompt

cleanup()
cleanup(): Promise<void>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:276

Cleanup agent resources.

Returns

Promise<void>

Inherited from

BaseAgent.cleanup

clearHistory()
protected clearHistory(): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:337

Returns

void

Inherited from

BaseAgent.clearHistory

complete()
protected complete(request): Promise<Result<CompletionResponse, AgentError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:303

Parameters
request

CompletionRequest

Returns

Promise<Result<CompletionResponse, AgentError>>

Inherited from

BaseAgent.complete

execute()
execute(task, options?): Promise<Result<TaskResult, AgentError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:218

Execute a task.

Parameters
task

Task

Task to execute

options?

Optional execution options (#3016/#3040). signal cancels the in-flight model call when the caller’s deadline wins a race; without it, the SDK keeps running to its own 10-minute timeout after the caller has already discarded the result.

signal?

AbortSignal

Returns

Promise<Result<TaskResult, AgentError>>

Result with TaskResult or AgentError

Inherited from

BaseAgent.execute

executeTask()
protected executeTask(task): Promise<Result<TaskResult, AgentError>>;

Defined in: packages/nexus-agents/src/agents/experts/architecture-expert.ts:77

Parameters
task

Task

Returns

Promise<Result<TaskResult, AgentError>>

Overrides

BaseAgent.executeTask

findResolutionForError()
protected findResolutionForError(errorMessage): ErrorResolution | undefined;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:393

Parameters
errorMessage

string

Returns

ErrorResolution | undefined

Inherited from

BaseAgent.findResolutionForError

flushMemory()
flushMemory(): Promise<Result<void, AgentMemoryError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:371

Returns

Promise<Result<void, AgentMemoryError>>

Inherited from

BaseAgent.flushMemory

getExpertOptions()
getExpertOptions(): Readonly<ArchitectureExpertOptions>;

Defined in: packages/nexus-agents/src/agents/experts/architecture-expert.ts:111

Returns

Readonly<ArchitectureExpertOptions>

getHistory()
protected getHistory(): Message[];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:334

Returns

Message[]

Inherited from

BaseAgent.getHistory

getMemoryState()
getMemoryState(): Readonly<AgentMemoryState> | null;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:364

Returns

Readonly<AgentMemoryState> | null

Inherited from

BaseAgent.getMemoryState

getPruningMetrics()
getPruningMetrics(): Readonly<ContextPruningMetrics>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:340

Returns

Readonly<ContextPruningMetrics>

Inherited from

BaseAgent.getPruningMetrics

getRelevantMemories()
getRelevantMemories(): readonly TypedMemoryEntry<MemoryType>[];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:367

Returns

readonly TypedMemoryEntry<MemoryType>[]

Inherited from

BaseAgent.getRelevantMemories

getTaskLearnings()
protected getTaskLearnings(taskType): readonly TaskLearning[];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:396

Parameters
taskType

string

Returns

readonly TaskLearning[]

Inherited from

BaseAgent.getTaskLearnings

handleMessage()
handleMessage(msg): Promise<Result<AgentResponse, AgentError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:261

Handle an inter-agent message and return a response.

Delivery semantics (#3222). This is a direct, awaited request/response call: the caller invokes it and holds the returned promise. It is NOT a queued or broadcast channel — that is the collaboration event bus (agents/collaboration/event-bus.ts), a fire-and-forget pub/sub with its own semantics. For this method specifically:

  • Ordering is the caller’s responsibility. Sequential awaits are handled in call order; concurrent calls carry no cross-message ordering guarantee.
  • Delivery is exactly the method invocation — there is no automatic retry or redelivery. A returned err(...) is the caller’s signal to decide whether to retry; the agent does not re-queue the message.
  • Errors surface as Result.err, not as a throw for expected conditions; the caller branches on the Result.
Parameters
msg

AgentMessage

Message to handle

Returns

Promise<Result<AgentResponse, AgentError>>

Result with AgentResponse, or AgentError on failure (not retried)

Inherited from

BaseAgent.handleMessage

hasCapability()
hasCapability(capability): boolean;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:292

Parameters
capability

AgentCapability

Returns

boolean

Inherited from

BaseAgent.hasCapability

initialize()
initialize(ctx): Promise<Result<void, AgentError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:205

Initialize the agent with context.

Parameters
ctx

AgentContext

Agent context

Returns

Promise<Result<void, AgentError>>

Result with void or AgentError

Inherited from

BaseAgent.initialize

isContextPruningEnabled()
isContextPruningEnabled(): boolean;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:358

Returns

boolean

Inherited from

BaseAgent.isContextPruningEnabled

isMemoryEnabled()
isMemoryEnabled(): boolean;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:361

Returns

boolean

Inherited from

BaseAgent.isMemoryEnabled

recordLearning()
protected recordLearning(learning): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:384

Parameters
learning

Omit<TaskLearning, "id" | "learnedAt">

Returns

void

Inherited from

BaseAgent.recordLearning

recordPattern()
protected recordPattern(p): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:387

Parameters
p

Omit<ExecutionPattern, "id" | "lastSeen" | "occurrences">

Returns

void

Inherited from

BaseAgent.recordPattern

recordResolution()
protected recordResolution(r): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:390

Parameters
r

Omit<ErrorResolution, "resolvedAt">

Returns

void

Inherited from

BaseAgent.recordResolution

transformError()
protected transformError(error, taskId): AgentError;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:299

Parameters
error

unknown

taskId

string

Returns

AgentError

Inherited from

BaseAgent.transformError


abstract BaseAgent

Defined in: packages/nexus-agents/src/agents/base-agent.ts:92

Abstract base class for all agents. Subclasses must implement executeTask and buildPrompt.

Extended by

Implements

Constructors

Constructor
new BaseAgent(options): BaseAgent;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:129

Parameters
options

BaseAgentOptions

Returns

BaseAgent

Properties

adapter
protected adapter: IModelAdapter | undefined;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:98

budgetTracker
protected readonly budgetTracker: ITokenBudgetTracker;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:97

capabilities
readonly capabilities: readonly AgentCapability[];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:95

Agent capabilities

Implementation of

IAgent.capabilities

config
protected config: AgentConfig | undefined;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:100

emitMessageEvents
protected readonly emitMessageEvents: boolean;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:107

eventBus
protected readonly eventBus: ICollaborationEventBus;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:106

history
protected history: Message[] = [];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:102

id
readonly id: string;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:93

Unique agent identifier

Implementation of

IAgent.id

logger
protected readonly logger: ILogger;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:99

maxTokens
protected readonly maxTokens: number;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:105

role
readonly role: AgentRole;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:94

Agent role

Implementation of

IAgent.role

sharedState
protected sharedState: Record<string, unknown> = {};

Defined in: packages/nexus-agents/src/agents/base-agent.ts:101

stateMachine
protected readonly stateMachine: AgentStateMachine;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:96

systemPrompt
protected readonly systemPrompt: string | undefined;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:103

temperature
protected readonly temperature: number;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:104

Accessors

state
Get Signature
get state(): AgentState;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:176

Current state

Returns

AgentState

Current state

Implementation of

IAgent.state

Methods

addContextItem()
protected addContextItem(
   content, 
   priority?, 
   category?
): Promise<void>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:344

Parameters
content

string

priority?

100 | 20 | 60 | 80 | 40

category?

"system" | "task" | "active"

Returns

Promise<void>

addToHistory()
protected addToHistory(message): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:331

Parameters
message

Message

Returns

void

buildPrompt()
abstract protected buildPrompt(task): Message[];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:297

Parameters
task

Task

Returns

Message[]

cleanup()
cleanup(): Promise<void>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:276

Cleanup agent resources.

Returns

Promise<void>

Implementation of

IAgent.cleanup

clearHistory()
protected clearHistory(): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:337

Returns

void

complete()
protected complete(request): Promise<Result<CompletionResponse, AgentError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:303

Parameters
request

CompletionRequest

Returns

Promise<Result<CompletionResponse, AgentError>>

execute()
execute(task, options?): Promise<Result<TaskResult, AgentError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:218

Execute a task.

Parameters
task

Task

Task to execute

options?

Optional execution options (#3016/#3040). signal cancels the in-flight model call when the caller’s deadline wins a race; without it, the SDK keeps running to its own 10-minute timeout after the caller has already discarded the result.

signal?

AbortSignal

Returns

Promise<Result<TaskResult, AgentError>>

Result with TaskResult or AgentError

Implementation of

IAgent.execute

executeTask()
abstract protected executeTask(task): Promise<Result<TaskResult, AgentError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:296

Parameters
task

Task

Returns

Promise<Result<TaskResult, AgentError>>

findResolutionForError()
protected findResolutionForError(errorMessage): ErrorResolution | undefined;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:393

Parameters
errorMessage

string

Returns

ErrorResolution | undefined

flushMemory()
flushMemory(): Promise<Result<void, AgentMemoryError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:371

Returns

Promise<Result<void, AgentMemoryError>>

getHistory()
protected getHistory(): Message[];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:334

Returns

Message[]

getMemoryState()
getMemoryState(): Readonly<AgentMemoryState> | null;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:364

Returns

Readonly<AgentMemoryState> | null

getPruningMetrics()
getPruningMetrics(): Readonly<ContextPruningMetrics>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:340

Returns

Readonly<ContextPruningMetrics>

getRelevantMemories()
getRelevantMemories(): readonly TypedMemoryEntry<MemoryType>[];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:367

Returns

readonly TypedMemoryEntry<MemoryType>[]

getTaskLearnings()
protected getTaskLearnings(taskType): readonly TaskLearning[];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:396

Parameters
taskType

string

Returns

readonly TaskLearning[]

handleMessage()
handleMessage(msg): Promise<Result<AgentResponse, AgentError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:261

Handle an inter-agent message and return a response.

Delivery semantics (#3222). This is a direct, awaited request/response call: the caller invokes it and holds the returned promise. It is NOT a queued or broadcast channel — that is the collaboration event bus (agents/collaboration/event-bus.ts), a fire-and-forget pub/sub with its own semantics. For this method specifically:

  • Ordering is the caller’s responsibility. Sequential awaits are handled in call order; concurrent calls carry no cross-message ordering guarantee.
  • Delivery is exactly the method invocation — there is no automatic retry or redelivery. A returned err(...) is the caller’s signal to decide whether to retry; the agent does not re-queue the message.
  • Errors surface as Result.err, not as a throw for expected conditions; the caller branches on the Result.
Parameters
msg

AgentMessage

Message to handle

Returns

Promise<Result<AgentResponse, AgentError>>

Result with AgentResponse, or AgentError on failure (not retried)

Implementation of

IAgent.handleMessage

hasCapability()
hasCapability(capability): boolean;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:292

Parameters
capability

AgentCapability

Returns

boolean

initialize()
initialize(ctx): Promise<Result<void, AgentError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:205

Initialize the agent with context.

Parameters
ctx

AgentContext

Agent context

Returns

Promise<Result<void, AgentError>>

Result with void or AgentError

Implementation of

IAgent.initialize

isContextPruningEnabled()
isContextPruningEnabled(): boolean;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:358

Returns

boolean

isMemoryEnabled()
isMemoryEnabled(): boolean;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:361

Returns

boolean

recordLearning()
protected recordLearning(learning): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:384

Parameters
learning

Omit<TaskLearning, "id" | "learnedAt">

Returns

void

recordPattern()
protected recordPattern(p): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:387

Parameters
p

Omit<ExecutionPattern, "id" | "lastSeen" | "occurrences">

Returns

void

recordResolution()
protected recordResolution(r): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:390

Parameters
r

Omit<ErrorResolution, "resolvedAt">

Returns

void

transformError()
protected transformError(error, taskId): AgentError;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:299

Parameters
error

unknown

taskId

string

Returns

AgentError


CodeExpert

Defined in: packages/nexus-agents/src/agents/experts/code-expert.ts:40

CodeExpert - Expert agent for code-related tasks.

Specialized in:

  • Code generation from specifications
  • Code refactoring and cleanup
  • Performance optimization
  • Bug detection and debugging

Extends

Constructors

Constructor
new CodeExpert(options?): CodeExpert;

Defined in: packages/nexus-agents/src/agents/experts/code-expert.ts:43

Parameters
options?

Partial<BaseAgentOptions> & { expertOptions?: CodeExpertOptions; } = {}

Returns

CodeExpert

Overrides

BaseAgent.constructor

Properties

adapter
protected adapter: IModelAdapter | undefined;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:98

Inherited from

BaseAgent.adapter

budgetTracker
protected readonly budgetTracker: ITokenBudgetTracker;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:97

Inherited from

BaseAgent.budgetTracker

capabilities
readonly capabilities: readonly AgentCapability[];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:95

Agent capabilities

Inherited from

BaseAgent.capabilities

config
protected config: AgentConfig | undefined;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:100

Inherited from

BaseAgent.config

emitMessageEvents
protected readonly emitMessageEvents: boolean;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:107

Inherited from

BaseAgent.emitMessageEvents

eventBus
protected readonly eventBus: ICollaborationEventBus;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:106

Inherited from

BaseAgent.eventBus

history
protected history: Message[] = [];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:102

Inherited from

BaseAgent.history

id
readonly id: string;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:93

Unique agent identifier

Inherited from

BaseAgent.id

logger
protected readonly logger: ILogger;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:99

Inherited from

BaseAgent.logger

maxTokens
protected readonly maxTokens: number;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:105

Inherited from

BaseAgent.maxTokens

role
readonly role: AgentRole;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:94

Agent role

Inherited from

BaseAgent.role

sharedState
protected sharedState: Record<string, unknown> = {};

Defined in: packages/nexus-agents/src/agents/base-agent.ts:101

Inherited from

BaseAgent.sharedState

stateMachine
protected readonly stateMachine: AgentStateMachine;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:96

Inherited from

BaseAgent.stateMachine

systemPrompt
protected readonly systemPrompt: string | undefined;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:103

Inherited from

BaseAgent.systemPrompt

temperature
protected readonly temperature: number;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:104

Inherited from

BaseAgent.temperature

Accessors

state
Get Signature
get state(): AgentState;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:176

Current state

Returns

AgentState

Current state

Inherited from

BaseAgent.state

Methods

addContextItem()
protected addContextItem(
   content, 
   priority?, 
   category?
): Promise<void>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:344

Parameters
content

string

priority?

100 | 20 | 60 | 80 | 40

category?

"system" | "task" | "active"

Returns

Promise<void>

Inherited from

BaseAgent.addContextItem

addToHistory()
protected addToHistory(message): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:331

Parameters
message

Message

Returns

void

Inherited from

BaseAgent.addToHistory

buildPrompt()
protected buildPrompt(task): Message[];

Defined in: packages/nexus-agents/src/agents/experts/code-expert.ts:76

Build prompt messages for the task.

Parameters
task

Task

Returns

Message[]

Overrides

BaseAgent.buildPrompt

cleanup()
cleanup(): Promise<void>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:276

Cleanup agent resources.

Returns

Promise<void>

Inherited from

BaseAgent.cleanup

clearHistory()
protected clearHistory(): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:337

Returns

void

Inherited from

BaseAgent.clearHistory

complete()
protected complete(request): Promise<Result<CompletionResponse, AgentError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:303

Parameters
request

CompletionRequest

Returns

Promise<Result<CompletionResponse, AgentError>>

Inherited from

BaseAgent.complete

execute()
execute(task, options?): Promise<Result<TaskResult, AgentError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:218

Execute a task.

Parameters
task

Task

Task to execute

options?

Optional execution options (#3016/#3040). signal cancels the in-flight model call when the caller’s deadline wins a race; without it, the SDK keeps running to its own 10-minute timeout after the caller has already discarded the result.

signal?

AbortSignal

Returns

Promise<Result<TaskResult, AgentError>>

Result with TaskResult or AgentError

Inherited from

BaseAgent.execute

executeTask()
protected executeTask(task): Promise<Result<TaskResult, AgentError>>;

Defined in: packages/nexus-agents/src/agents/experts/code-expert.ts:54

Execute a code-related task.

Parameters
task

Task

Returns

Promise<Result<TaskResult, AgentError>>

Overrides

BaseAgent.executeTask

findResolutionForError()
protected findResolutionForError(errorMessage): ErrorResolution | undefined;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:393

Parameters
errorMessage

string

Returns

ErrorResolution | undefined

Inherited from

BaseAgent.findResolutionForError

flushMemory()
flushMemory(): Promise<Result<void, AgentMemoryError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:371

Returns

Promise<Result<void, AgentMemoryError>>

Inherited from

BaseAgent.flushMemory

getExpertOptions()
getExpertOptions(): Readonly<CodeExpertOptions>;

Defined in: packages/nexus-agents/src/agents/experts/code-expert.ts:95

Get the expert options.

Returns

Readonly<CodeExpertOptions>

getHistory()
protected getHistory(): Message[];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:334

Returns

Message[]

Inherited from

BaseAgent.getHistory

getMemoryState()
getMemoryState(): Readonly<AgentMemoryState> | null;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:364

Returns

Readonly<AgentMemoryState> | null

Inherited from

BaseAgent.getMemoryState

getPruningMetrics()
getPruningMetrics(): Readonly<ContextPruningMetrics>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:340

Returns

Readonly<ContextPruningMetrics>

Inherited from

BaseAgent.getPruningMetrics

getRelevantMemories()
getRelevantMemories(): readonly TypedMemoryEntry<MemoryType>[];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:367

Returns

readonly TypedMemoryEntry<MemoryType>[]

Inherited from

BaseAgent.getRelevantMemories

getTaskLearnings()
protected getTaskLearnings(taskType): readonly TaskLearning[];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:396

Parameters
taskType

string

Returns

readonly TaskLearning[]

Inherited from

BaseAgent.getTaskLearnings

handleMessage()
handleMessage(msg): Promise<Result<AgentResponse, AgentError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:261

Handle an inter-agent message and return a response.

Delivery semantics (#3222). This is a direct, awaited request/response call: the caller invokes it and holds the returned promise. It is NOT a queued or broadcast channel — that is the collaboration event bus (agents/collaboration/event-bus.ts), a fire-and-forget pub/sub with its own semantics. For this method specifically:

  • Ordering is the caller’s responsibility. Sequential awaits are handled in call order; concurrent calls carry no cross-message ordering guarantee.
  • Delivery is exactly the method invocation — there is no automatic retry or redelivery. A returned err(...) is the caller’s signal to decide whether to retry; the agent does not re-queue the message.
  • Errors surface as Result.err, not as a throw for expected conditions; the caller branches on the Result.
Parameters
msg

AgentMessage

Message to handle

Returns

Promise<Result<AgentResponse, AgentError>>

Result with AgentResponse, or AgentError on failure (not retried)

Inherited from

BaseAgent.handleMessage

hasCapability()
hasCapability(capability): boolean;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:292

Parameters
capability

AgentCapability

Returns

boolean

Inherited from

BaseAgent.hasCapability

initialize()
initialize(ctx): Promise<Result<void, AgentError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:205

Initialize the agent with context.

Parameters
ctx

AgentContext

Agent context

Returns

Promise<Result<void, AgentError>>

Result with void or AgentError

Inherited from

BaseAgent.initialize

isContextPruningEnabled()
isContextPruningEnabled(): boolean;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:358

Returns

boolean

Inherited from

BaseAgent.isContextPruningEnabled

isMemoryEnabled()
isMemoryEnabled(): boolean;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:361

Returns

boolean

Inherited from

BaseAgent.isMemoryEnabled

recordLearning()
protected recordLearning(learning): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:384

Parameters
learning

Omit<TaskLearning, "id" | "learnedAt">

Returns

void

Inherited from

BaseAgent.recordLearning

recordPattern()
protected recordPattern(p): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:387

Parameters
p

Omit<ExecutionPattern, "id" | "lastSeen" | "occurrences">

Returns

void

Inherited from

BaseAgent.recordPattern

recordResolution()
protected recordResolution(r): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:390

Parameters
r

Omit<ErrorResolution, "resolvedAt">

Returns

void

Inherited from

BaseAgent.recordResolution

transformError()
protected transformError(error, taskId): AgentError;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:299

Parameters
error

unknown

taskId

string

Returns

AgentError

Inherited from

BaseAgent.transformError


CollaborationSession

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-session.ts:49

Manages a collaboration session between multiple experts.

Constructors

Constructor
new CollaborationSession(options?): CollaborationSession;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-session.ts:59

Parameters
options?

CollaborationSessionOptions = {}

Returns

CollaborationSession

Methods

addEventListener()
addEventListener(listener): void;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-session.ts:359

Parameters
listener

(event) => void

Returns

void

cancel()
cancel(reason): void;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-session.ts:349

Parameters
reason

string

Returns

void

finalize()
finalize(): Result<CollaborationResult, AgentError>;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-session.ts:298

Finalizes the session and returns aggregated result.

Returns

Result<CollaborationResult, AgentError>

getSessionId()
getSessionId(): string | null;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-session.ts:247

Returns

string | null

getStatus()
getStatus(): SessionState | null;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-session.ts:244

Returns

SessionState | null

getTaskAssignments()
getTaskAssignments(): TaskAssignmentMessage[];

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-session.ts:291

Gets task assignments for experts based on pattern.

Returns

TaskAssignmentMessage[]

markExpertFailed()
markExpertFailed(expertId, error): Result<void, AgentError>;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-session.ts:252

Marks an expert as failed.

Parameters
expertId

string

error

string

Returns

Result<void, AgentError>

removeEventListener()
removeEventListener(listener): void;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-session.ts:370

Parameters
listener

(event) => void

Returns

void

requestReview()
requestReview(
   fromExpert, 
   toExpert, 
   artifact
): Result<void, AgentError>;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-session.ts:150

Requests a review from one expert to another.

Parameters
fromExpert

string

toExpert

string

artifact

unknown

Returns

Result<void, AgentError>

start()
start(config): Result<string, AgentError>;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-session.ts:68

Starts a new collaboration session.

Parameters
config

CollaborationConfig

Returns

Result<string, AgentError>

submitResult()
submitResult(expertId, result): Result<void, AgentError>;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-session.ts:118

Submits a result from an expert.

Parameters
expertId

string

result

TaskResult

Returns

Result<void, AgentError>

submitReview()
submitReview(
   reviewerId, 
   requesterId, 
   approved, 
   feedback
): Result<void, AgentError>;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-session.ts:171

Submits a review response.

Parameters
reviewerId

string

requesterId

string

approved

boolean

feedback

string

Returns

Result<void, AgentError>

vote()
vote(
   expertId, 
   decision, 
   reasoning
): Result<void, AgentError>;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-session.ts:201

Submits a vote for consensus protocol.

Parameters
expertId

string

decision

"approve" | "reject" | "abstain"

reasoning

string

Returns

Result<void, AgentError>


ConsensusProtocol

Defined in: packages/nexus-agents/src/agents/collaboration/consensus-protocol.ts:22

Consensus collaboration protocol.

Implements

Constructors

Constructor
new ConsensusProtocol(options?): ConsensusProtocol;

Defined in: packages/nexus-agents/src/agents/collaboration/consensus-protocol.ts:29

Parameters
options?

ProtocolOptions = {}

Returns

ConsensusProtocol

Properties

cancelled
protected cancelled: boolean = false;

Defined in: packages/nexus-agents/src/agents/collaboration/consensus-protocol.ts:26

logger
protected readonly logger: ILogger;

Defined in: packages/nexus-agents/src/agents/collaboration/consensus-protocol.ts:24

options
protected readonly options: ProtocolOptions;

Defined in: packages/nexus-agents/src/agents/collaboration/consensus-protocol.ts:27

pattern
readonly pattern: "consensus";

Defined in: packages/nexus-agents/src/agents/collaboration/consensus-protocol.ts:23

Implementation of

ICollaborationProtocol.pattern

session
protected session: CollaborationSession | null = null;

Defined in: packages/nexus-agents/src/agents/collaboration/consensus-protocol.ts:25

Methods

cancel()
cancel(reason): void;

Defined in: packages/nexus-agents/src/agents/collaboration/consensus-protocol.ts:34

Parameters
reason

string

Returns

void

Implementation of

ICollaborationProtocol.cancel

execute()
execute(config, agents): Promise<Result<CollaborationResult, AgentError>>;

Defined in: packages/nexus-agents/src/agents/collaboration/consensus-protocol.ts:40

Parameters
config

CollaborationConfig

agents

Map<string, IAgent>

Returns

Promise<Result<CollaborationResult, AgentError>>

Implementation of

ICollaborationProtocol.execute


ContextManager

Defined in: packages/nexus-agents/src/agents/context-manager.ts:58

Manages context window for agents with token budget enforcement.

Constructors

Constructor
new ContextManager(config): ContextManager;

Defined in: packages/nexus-agents/src/agents/context-manager.ts:71

Parameters
config

ContextManagerConfig

Returns

ContextManager

Methods

add()
add(item): Promise<Result<ContextItem, ValidationError>>;

Defined in: packages/nexus-agents/src/agents/context-manager.ts:90

Add an item to the context. Returns Result with the added item or error.

Parameters
item

Omit<ContextItem, "tokenCount" | "addedAt">

Returns

Promise<Result<ContextItem, ValidationError>>

buildMessages()
buildMessages(): Message[];

Defined in: packages/nexus-agents/src/agents/context-manager.ts:247

Build messages array from context items for model requests.

Returns

Message[]

canAdd()
canAdd(content, category): Promise<boolean>;

Defined in: packages/nexus-agents/src/agents/context-manager.ts:221

Check if an item with the given content can be added to a category.

Parameters
content

string

category

"system" | "task" | "active"

Returns

Promise<boolean>

clear()
clear(): void;

Defined in: packages/nexus-agents/src/agents/context-manager.ts:280

Clear all items from the context.

Returns

void

clearCategory()
clearCategory(category): number;

Defined in: packages/nexus-agents/src/agents/context-manager.ts:288

Clear items from a specific category. Returns number of items removed.

Parameters
category

"system" | "task" | "active"

Returns

number

countTokens()
countTokens(text): Promise<number>;

Defined in: packages/nexus-agents/src/agents/context-manager.ts:308

Count tokens in text using adapter or fallback estimation.

Parameters
text

string

Returns

Promise<number>

get()
get(id): ContextItem | undefined;

Defined in: packages/nexus-agents/src/agents/context-manager.ts:216

Get an item by ID.

Parameters
id

string

Returns

ContextItem | undefined

getAllItems()
getAllItems(): ContextItem[];

Defined in: packages/nexus-agents/src/agents/context-manager.ts:242

Get all items sorted by priority (desc) then addedAt (asc).

Returns

ContextItem[]

getByCategory()
getByCategory(category): ContextItem[];

Defined in: packages/nexus-agents/src/agents/context-manager.ts:237

Get all items in a category, sorted by priority (desc) then addedAt (asc).

Parameters
category

"system" | "task" | "active"

Returns

ContextItem[]

getRemainingTokens()
getRemainingTokens(category): number;

Defined in: packages/nexus-agents/src/agents/context-manager.ts:267

Get remaining tokens available in a category.

Parameters
category

"system" | "task" | "active"

Returns

number

getStats()
getStats(): ContextStats;

Defined in: packages/nexus-agents/src/agents/context-manager.ts:257

Get current context statistics.

Returns

ContextStats

getSystemPrompt()
getSystemPrompt(): string | undefined;

Defined in: packages/nexus-agents/src/agents/context-manager.ts:252

Get the system prompt from system category items.

Returns

string | undefined

getTotalRemainingTokens()
getTotalRemainingTokens(): number;

Defined in: packages/nexus-agents/src/agents/context-manager.ts:274

Get total remaining tokens across all categories.

Returns

number

remove()
remove(id): boolean;

Defined in: packages/nexus-agents/src/agents/context-manager.ts:200

Remove an item from the context. Returns true if removed, false if not found.

Parameters
id

string

Returns

boolean


ContextPruner

Defined in: packages/nexus-agents/src/agents/context-pruner.ts:96

Handles context pruning with multiple strategies.

Constructors

Constructor
new ContextPruner(config): ContextPruner;

Defined in: packages/nexus-agents/src/agents/context-pruner.ts:105

Parameters
config

ContextPrunerConfig

Returns

ContextPruner

Methods

estimateFreeableTokens()
estimateFreeableTokens(categories): number;

Defined in: packages/nexus-agents/src/agents/context-pruner.ts:193

Estimate tokens that can be freed from specified categories.

Parameters
categories

("system" | "task" | "active")[]

Returns

number

getPruneCandidates()
getPruneCandidates(categories): ContextItem[];

Defined in: packages/nexus-agents/src/agents/context-pruner.ts:184

Get candidates for pruning from specified categories.

Parameters
categories

("system" | "task" | "active")[]

Returns

ContextItem[]

prune()
prune(options?): Promise<Result<PruneResult, ValidationError>>;

Defined in: packages/nexus-agents/src/agents/context-pruner.ts:132

Prune context to free tokens or reach target capacity.

Parameters
options?

PruneOptions = {}

Returns

Promise<Result<PruneResult, ValidationError>>

pruneCategory()
pruneCategory(category, targetTokens): Promise<Result<PruneResult, ValidationError>>;

Defined in: packages/nexus-agents/src/agents/context-pruner.ts:176

Prune items from a specific category.

Parameters
category

"system" | "task" | "active"

targetTokens

number

Returns

Promise<Result<PruneResult, ValidationError>>

shouldPrune()
shouldPrune(): boolean;

Defined in: packages/nexus-agents/src/agents/context-pruner.ts:127

Check if pruning should be triggered based on usage threshold.

Returns

boolean


DocumentationExpert

Defined in: packages/nexus-agents/src/agents/experts/documentation-expert.ts:52

DocumentationExpert - Expert agent for documentation-related tasks.

Extends

Constructors

Constructor
new DocumentationExpert(options?): DocumentationExpert;

Defined in: packages/nexus-agents/src/agents/experts/documentation-expert.ts:55

Parameters
options?

Partial<BaseAgentOptions> & { expertOptions?: DocumentationExpertOptions; } = {}

Returns

DocumentationExpert

Overrides

BaseAgent.constructor

Properties

adapter
protected adapter: IModelAdapter | undefined;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:98

Inherited from

BaseAgent.adapter

budgetTracker
protected readonly budgetTracker: ITokenBudgetTracker;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:97

Inherited from

BaseAgent.budgetTracker

capabilities
readonly capabilities: readonly AgentCapability[];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:95

Agent capabilities

Inherited from

BaseAgent.capabilities

config
protected config: AgentConfig | undefined;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:100

Inherited from

BaseAgent.config

emitMessageEvents
protected readonly emitMessageEvents: boolean;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:107

Inherited from

BaseAgent.emitMessageEvents

eventBus
protected readonly eventBus: ICollaborationEventBus;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:106

Inherited from

BaseAgent.eventBus

history
protected history: Message[] = [];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:102

Inherited from

BaseAgent.history

id
readonly id: string;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:93

Unique agent identifier

Inherited from

BaseAgent.id

logger
protected readonly logger: ILogger;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:99

Inherited from

BaseAgent.logger

maxTokens
protected readonly maxTokens: number;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:105

Inherited from

BaseAgent.maxTokens

role
readonly role: AgentRole;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:94

Agent role

Inherited from

BaseAgent.role

sharedState
protected sharedState: Record<string, unknown> = {};

Defined in: packages/nexus-agents/src/agents/base-agent.ts:101

Inherited from

BaseAgent.sharedState

stateMachine
protected readonly stateMachine: AgentStateMachine;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:96

Inherited from

BaseAgent.stateMachine

systemPrompt
protected readonly systemPrompt: string | undefined;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:103

Inherited from

BaseAgent.systemPrompt

temperature
protected readonly temperature: number;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:104

Inherited from

BaseAgent.temperature

Accessors

state
Get Signature
get state(): AgentState;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:176

Current state

Returns

AgentState

Current state

Inherited from

BaseAgent.state

Methods

addContextItem()
protected addContextItem(
   content, 
   priority?, 
   category?
): Promise<void>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:344

Parameters
content

string

priority?

100 | 20 | 60 | 80 | 40

category?

"system" | "task" | "active"

Returns

Promise<void>

Inherited from

BaseAgent.addContextItem

addToHistory()
protected addToHistory(message): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:331

Parameters
message

Message

Returns

void

Inherited from

BaseAgent.addToHistory

buildPrompt()
protected buildPrompt(task): Message[];

Defined in: packages/nexus-agents/src/agents/experts/documentation-expert.ts:83

Parameters
task

Task

Returns

Message[]

Overrides

BaseAgent.buildPrompt

cleanup()
cleanup(): Promise<void>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:276

Cleanup agent resources.

Returns

Promise<void>

Inherited from

BaseAgent.cleanup

clearHistory()
protected clearHistory(): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:337

Returns

void

Inherited from

BaseAgent.clearHistory

complete()
protected complete(request): Promise<Result<CompletionResponse, AgentError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:303

Parameters
request

CompletionRequest

Returns

Promise<Result<CompletionResponse, AgentError>>

Inherited from

BaseAgent.complete

execute()
execute(task, options?): Promise<Result<TaskResult, AgentError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:218

Execute a task.

Parameters
task

Task

Task to execute

options?

Optional execution options (#3016/#3040). signal cancels the in-flight model call when the caller’s deadline wins a race; without it, the SDK keeps running to its own 10-minute timeout after the caller has already discarded the result.

signal?

AbortSignal

Returns

Promise<Result<TaskResult, AgentError>>

Result with TaskResult or AgentError

Inherited from

BaseAgent.execute

executeTask()
protected executeTask(task): Promise<Result<TaskResult, AgentError>>;

Defined in: packages/nexus-agents/src/agents/experts/documentation-expert.ts:65

Parameters
task

Task

Returns

Promise<Result<TaskResult, AgentError>>

Overrides

BaseAgent.executeTask

findResolutionForError()
protected findResolutionForError(errorMessage): ErrorResolution | undefined;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:393

Parameters
errorMessage

string

Returns

ErrorResolution | undefined

Inherited from

BaseAgent.findResolutionForError

flushMemory()
flushMemory(): Promise<Result<void, AgentMemoryError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:371

Returns

Promise<Result<void, AgentMemoryError>>

Inherited from

BaseAgent.flushMemory

getExpertOptions()
getExpertOptions(): Readonly<DocumentationExpertOptions>;

Defined in: packages/nexus-agents/src/agents/experts/documentation-expert.ts:99

Returns

Readonly<DocumentationExpertOptions>

getHistory()
protected getHistory(): Message[];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:334

Returns

Message[]

Inherited from

BaseAgent.getHistory

getMemoryState()
getMemoryState(): Readonly<AgentMemoryState> | null;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:364

Returns

Readonly<AgentMemoryState> | null

Inherited from

BaseAgent.getMemoryState

getPruningMetrics()
getPruningMetrics(): Readonly<ContextPruningMetrics>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:340

Returns

Readonly<ContextPruningMetrics>

Inherited from

BaseAgent.getPruningMetrics

getRelevantMemories()
getRelevantMemories(): readonly TypedMemoryEntry<MemoryType>[];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:367

Returns

readonly TypedMemoryEntry<MemoryType>[]

Inherited from

BaseAgent.getRelevantMemories

getTaskLearnings()
protected getTaskLearnings(taskType): readonly TaskLearning[];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:396

Parameters
taskType

string

Returns

readonly TaskLearning[]

Inherited from

BaseAgent.getTaskLearnings

handleMessage()
handleMessage(msg): Promise<Result<AgentResponse, AgentError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:261

Handle an inter-agent message and return a response.

Delivery semantics (#3222). This is a direct, awaited request/response call: the caller invokes it and holds the returned promise. It is NOT a queued or broadcast channel — that is the collaboration event bus (agents/collaboration/event-bus.ts), a fire-and-forget pub/sub with its own semantics. For this method specifically:

  • Ordering is the caller’s responsibility. Sequential awaits are handled in call order; concurrent calls carry no cross-message ordering guarantee.
  • Delivery is exactly the method invocation — there is no automatic retry or redelivery. A returned err(...) is the caller’s signal to decide whether to retry; the agent does not re-queue the message.
  • Errors surface as Result.err, not as a throw for expected conditions; the caller branches on the Result.
Parameters
msg

AgentMessage

Message to handle

Returns

Promise<Result<AgentResponse, AgentError>>

Result with AgentResponse, or AgentError on failure (not retried)

Inherited from

BaseAgent.handleMessage

hasCapability()
hasCapability(capability): boolean;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:292

Parameters
capability

AgentCapability

Returns

boolean

Inherited from

BaseAgent.hasCapability

initialize()
initialize(ctx): Promise<Result<void, AgentError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:205

Initialize the agent with context.

Parameters
ctx

AgentContext

Agent context

Returns

Promise<Result<void, AgentError>>

Result with void or AgentError

Inherited from

BaseAgent.initialize

isContextPruningEnabled()
isContextPruningEnabled(): boolean;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:358

Returns

boolean

Inherited from

BaseAgent.isContextPruningEnabled

isMemoryEnabled()
isMemoryEnabled(): boolean;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:361

Returns

boolean

Inherited from

BaseAgent.isMemoryEnabled

recordLearning()
protected recordLearning(learning): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:384

Parameters
learning

Omit<TaskLearning, "id" | "learnedAt">

Returns

void

Inherited from

BaseAgent.recordLearning

recordPattern()
protected recordPattern(p): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:387

Parameters
p

Omit<ExecutionPattern, "id" | "lastSeen" | "occurrences">

Returns

void

Inherited from

BaseAgent.recordPattern

recordResolution()
protected recordResolution(r): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:390

Parameters
r

Omit<ErrorResolution, "resolvedAt">

Returns

void

Inherited from

BaseAgent.recordResolution

transformError()
protected transformError(error, taskId): AgentError;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:299

Parameters
error

unknown

taskId

string

Returns

AgentError

Inherited from

BaseAgent.transformError


Expert

Defined in: packages/nexus-agents/src/agents/experts/expert-agent.ts:17

Expert agent extending SimpleAgent with configuration-based setup.

Extends

Extended by

Constructors

Constructor
new Expert(options, config): Expert;

Defined in: packages/nexus-agents/src/agents/experts/expert-agent.ts:20

Parameters
options

BaseAgentOptions

config

ExpertConfig

Returns

Expert

Overrides

SimpleAgent.constructor

Properties

adapter
protected adapter: IModelAdapter | undefined;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:98

Inherited from

SimpleAgent.adapter

budgetTracker
protected readonly budgetTracker: ITokenBudgetTracker;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:97

Inherited from

SimpleAgent.budgetTracker

capabilities
readonly capabilities: readonly AgentCapability[];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:95

Agent capabilities

Inherited from

SimpleAgent.capabilities

config
protected config: AgentConfig | undefined;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:100

Inherited from

SimpleAgent.config

emitMessageEvents
protected readonly emitMessageEvents: boolean;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:107

Inherited from

SimpleAgent.emitMessageEvents

eventBus
protected readonly eventBus: ICollaborationEventBus;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:106

Inherited from

SimpleAgent.eventBus

expertConfig
readonly expertConfig: ExpertConfig;

Defined in: packages/nexus-agents/src/agents/experts/expert-agent.ts:18

history
protected history: Message[] = [];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:102

Inherited from

SimpleAgent.history

id
readonly id: string;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:93

Unique agent identifier

Inherited from

SimpleAgent.id

logger
protected readonly logger: ILogger;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:99

Inherited from

SimpleAgent.logger

maxTokens
protected readonly maxTokens: number;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:105

Inherited from

SimpleAgent.maxTokens

role
readonly role: AgentRole;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:94

Agent role

Inherited from

SimpleAgent.role

sharedState
protected sharedState: Record<string, unknown> = {};

Defined in: packages/nexus-agents/src/agents/base-agent.ts:101

Inherited from

SimpleAgent.sharedState

stateMachine
protected readonly stateMachine: AgentStateMachine;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:96

Inherited from

SimpleAgent.stateMachine

systemPrompt
protected readonly systemPrompt: string | undefined;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:103

Inherited from

SimpleAgent.systemPrompt

temperature
protected readonly temperature: number;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:104

Inherited from

SimpleAgent.temperature

Accessors

metadata
Get Signature
get metadata(): Record<string, unknown> | undefined;

Defined in: packages/nexus-agents/src/agents/experts/expert-agent.ts:35

Get the expert’s metadata.

Returns

Record<string, unknown> | undefined

name
Get Signature
get name(): string;

Defined in: packages/nexus-agents/src/agents/experts/expert-agent.ts:28

Get the expert’s name.

Returns

string

state
Get Signature
get state(): AgentState;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:176

Current state

Returns

AgentState

Current state

Inherited from

SimpleAgent.state

Methods

addContextItem()
protected addContextItem(
   content, 
   priority?, 
   category?
): Promise<void>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:344

Parameters
content

string

priority?

100 | 20 | 60 | 80 | 40

category?

"system" | "task" | "active"

Returns

Promise<void>

Inherited from

SimpleAgent.addContextItem

addToHistory()
protected addToHistory(message): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:331

Parameters
message

Message

Returns

void

Inherited from

SimpleAgent.addToHistory

buildPrompt()
protected buildPrompt(task): Message[];

Defined in: packages/nexus-agents/src/agents/simple-agent.ts:105

Build prompt messages from a task.

Parameters
task

Task

Returns

Message[]

Inherited from

SimpleAgent.buildPrompt

cleanup()
cleanup(): Promise<void>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:276

Cleanup agent resources.

Returns

Promise<void>

Inherited from

SimpleAgent.cleanup

clearHistory()
protected clearHistory(): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:337

Returns

void

Inherited from

SimpleAgent.clearHistory

complete()
protected complete(request): Promise<Result<CompletionResponse, AgentError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:303

Parameters
request

CompletionRequest

Returns

Promise<Result<CompletionResponse, AgentError>>

Inherited from

SimpleAgent.complete

execute()
execute(task, options?): Promise<Result<TaskResult, AgentError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:218

Execute a task.

Parameters
task

Task

Task to execute

options?

Optional execution options (#3016/#3040). signal cancels the in-flight model call when the caller’s deadline wins a race; without it, the SDK keeps running to its own 10-minute timeout after the caller has already discarded the result.

signal?

AbortSignal

Returns

Promise<Result<TaskResult, AgentError>>

Result with TaskResult or AgentError

Inherited from

SimpleAgent.execute

executeTask()
protected executeTask(task): Promise<Result<TaskResult, AgentError>>;

Defined in: packages/nexus-agents/src/agents/simple-agent.ts:21

Execute a task by sending it to the model.

Parameters
task

Task

Returns

Promise<Result<TaskResult, AgentError>>

Inherited from

SimpleAgent.executeTask

findResolutionForError()
protected findResolutionForError(errorMessage): ErrorResolution | undefined;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:393

Parameters
errorMessage

string

Returns

ErrorResolution | undefined

Inherited from

SimpleAgent.findResolutionForError

flushMemory()
flushMemory(): Promise<Result<void, AgentMemoryError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:371

Returns

Promise<Result<void, AgentMemoryError>>

Inherited from

SimpleAgent.flushMemory

getHistory()
protected getHistory(): Message[];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:334

Returns

Message[]

Inherited from

SimpleAgent.getHistory

getMemoryState()
getMemoryState(): Readonly<AgentMemoryState> | null;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:364

Returns

Readonly<AgentMemoryState> | null

Inherited from

SimpleAgent.getMemoryState

getPruningMetrics()
getPruningMetrics(): Readonly<ContextPruningMetrics>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:340

Returns

Readonly<ContextPruningMetrics>

Inherited from

SimpleAgent.getPruningMetrics

getRelevantMemories()
getRelevantMemories(): readonly TypedMemoryEntry<MemoryType>[];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:367

Returns

readonly TypedMemoryEntry<MemoryType>[]

Inherited from

SimpleAgent.getRelevantMemories

getTaskLearnings()
protected getTaskLearnings(taskType): readonly TaskLearning[];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:396

Parameters
taskType

string

Returns

readonly TaskLearning[]

Inherited from

SimpleAgent.getTaskLearnings

handleMessage()
handleMessage(msg): Promise<Result<AgentResponse, AgentError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:261

Handle an inter-agent message and return a response.

Delivery semantics (#3222). This is a direct, awaited request/response call: the caller invokes it and holds the returned promise. It is NOT a queued or broadcast channel — that is the collaboration event bus (agents/collaboration/event-bus.ts), a fire-and-forget pub/sub with its own semantics. For this method specifically:

  • Ordering is the caller’s responsibility. Sequential awaits are handled in call order; concurrent calls carry no cross-message ordering guarantee.
  • Delivery is exactly the method invocation — there is no automatic retry or redelivery. A returned err(...) is the caller’s signal to decide whether to retry; the agent does not re-queue the message.
  • Errors surface as Result.err, not as a throw for expected conditions; the caller branches on the Result.
Parameters
msg

AgentMessage

Message to handle

Returns

Promise<Result<AgentResponse, AgentError>>

Result with AgentResponse, or AgentError on failure (not retried)

Inherited from

SimpleAgent.handleMessage

hasCapability()
hasCapability(capability): boolean;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:292

Parameters
capability

AgentCapability

Returns

boolean

Inherited from

SimpleAgent.hasCapability

initialize()
initialize(ctx): Promise<Result<void, AgentError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:205

Initialize the agent with context.

Parameters
ctx

AgentContext

Agent context

Returns

Promise<Result<void, AgentError>>

Result with void or AgentError

Inherited from

SimpleAgent.initialize

isContextPruningEnabled()
isContextPruningEnabled(): boolean;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:358

Returns

boolean

Inherited from

SimpleAgent.isContextPruningEnabled

isMemoryEnabled()
isMemoryEnabled(): boolean;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:361

Returns

boolean

Inherited from

SimpleAgent.isMemoryEnabled

recordLearning()
protected recordLearning(learning): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:384

Parameters
learning

Omit<TaskLearning, "id" | "learnedAt">

Returns

void

Inherited from

SimpleAgent.recordLearning

recordPattern()
protected recordPattern(p): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:387

Parameters
p

Omit<ExecutionPattern, "id" | "lastSeen" | "occurrences">

Returns

void

Inherited from

SimpleAgent.recordPattern

recordResolution()
protected recordResolution(r): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:390

Parameters
r

Omit<ErrorResolution, "resolvedAt">

Returns

void

Inherited from

SimpleAgent.recordResolution

transformError()
protected transformError(error, taskId): AgentError;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:299

Parameters
error

unknown

taskId

string

Returns

AgentError

Inherited from

SimpleAgent.transformError


ExpertRegistry

Defined in: packages/nexus-agents/src/agents/experts/expert-registry.ts:68

Singleton registry for managing expert agents.

Provides thread-safe registration and lookup of experts. Supports querying by ID, role, and capabilities.

Implements IRegistry<Expert, RegistryError> for unified registry API.

Implements

Accessors

isEmpty
Get Signature
get isEmpty(): boolean;

Defined in: packages/nexus-agents/src/agents/experts/expert-registry.ts:319

Check if the registry is empty.

Returns

boolean

Implementation of
IRegistry.isEmpty
size
Get Signature
get size(): number;

Defined in: packages/nexus-agents/src/agents/experts/expert-registry.ts:312

Get the number of registered experts.

Returns

number

Implementation of
IRegistry.size

Methods

clear()
clear(): void;

Defined in: packages/nexus-agents/src/agents/experts/expert-registry.ts:326

Clear all registered experts.

Returns

void

Implementation of
IRegistry.clear
findBestMatch()
findBestMatch(requiredCapabilities): Result<Expert, RegistryError>;

Defined in: packages/nexus-agents/src/agents/experts/expert-registry.ts:363

Find the best expert for a set of required capabilities.

Returns the expert that matches the most capabilities.

Parameters
requiredCapabilities

AgentCapability[]

Capabilities needed

Returns

Result<Expert, RegistryError>

Result with best Expert or RegistryError if none found

get()
get(id): Result<Expert, RegistryError>;

Defined in: packages/nexus-agents/src/agents/experts/expert-registry.ts:163

Get an expert by ID.

Parameters
id

string

Expert ID to retrieve

Returns

Result<Expert, RegistryError>

Result with Expert or RegistryError

Implementation of
IRegistry.get
getAll()
getAll(): Expert[];

Defined in: packages/nexus-agents/src/agents/experts/expert-registry.ts:274

Get all registered experts. IRegistry interface method.

Returns

Expert[]

Array of all registered experts

Implementation of
IRegistry.getAll
getAllIds()
getAllIds(): string[];

Defined in: packages/nexus-agents/src/agents/experts/expert-registry.ts:284

Get all registered expert IDs. IRegistry interface method.

Returns

string[]

Array of all registered expert IDs

Implementation of
IRegistry.getAllIds
getByCapability()
getByCapability(capability): Expert[];

Defined in: packages/nexus-agents/src/agents/experts/expert-registry.ts:198

Get experts by capability.

Returns all experts that have the specified capability.

Parameters
capability

AgentCapability

Capability to search for

Returns

Expert[]

Array of matching experts

getByRole()
getByRole(role): Expert[];

Defined in: packages/nexus-agents/src/agents/experts/expert-registry.ts:210

Get experts by role.

Parameters
role

string

Role to search for

Returns

Expert[]

Array of matching experts

getStats()
getStats(): RegistryStats;

Defined in: packages/nexus-agents/src/agents/experts/expert-registry.ts:334

Get statistics about the registry. Returns IRegistryStats-compatible stats with domain-specific extensions.

Returns

RegistryStats

Implementation of
IRegistry.getStats
has()
has(id): boolean;

Defined in: packages/nexus-agents/src/agents/experts/expert-registry.ts:186

Check if an expert is registered.

Parameters
id

string

Expert ID to check

Returns

boolean

True if expert is registered

Implementation of
IRegistry.has
query()
query(predicate): Expert[];

Defined in: packages/nexus-agents/src/agents/experts/expert-registry.ts:260

Query experts with predicate function. IRegistry interface method.

Parameters
predicate

(item) => boolean

Function to test each expert

Returns

Expert[]

Array of matching experts

Implementation of
IRegistry.query
queryWithOptions()
queryWithOptions(options): Expert[];

Defined in: packages/nexus-agents/src/agents/experts/expert-registry.ts:221

Query experts with multiple criteria. Domain-specific query with structured options.

Parameters
options

QueryOptions

Query options

Returns

Expert[]

Array of matching experts

register()
register(expert, options?): Result<void, RegistryError>;

Defined in: packages/nexus-agents/src/agents/experts/expert-registry.ts:101

Register an expert in the registry.

Parameters
expert

Expert

Expert to register

options?

ExpertRegisterOptions

Registration options

Returns

Result<void, RegistryError>

Result with void or RegistryError

Implementation of
IRegistry.register
registerMany()
registerMany(experts, options?): Result<void, RegistryError>;

Defined in: packages/nexus-agents/src/agents/experts/expert-registry.ts:126

Register multiple experts.

Parameters
experts

Expert[]

Experts to register

options?

ExpertRegisterOptions

Registration options

Returns

Result<void, RegistryError>

Result with void or first RegistryError

search(searchTerm): Expert[];

Defined in: packages/nexus-agents/src/agents/experts/expert-registry.ts:297

Search experts by text query. IRegistry interface method.

Searches expert ID, name, role, and capabilities.

Parameters
searchTerm

string

Search term to match

Returns

Expert[]

Array of matching experts

Implementation of
IRegistry.search
unregister()
unregister(id): Result<Expert, RegistryError>;

Defined in: packages/nexus-agents/src/agents/experts/expert-registry.ts:142

Unregister an expert by ID.

Parameters
id

string

Expert ID to unregister

Returns

Result<Expert, RegistryError>

Result with the removed Expert or RegistryError

Implementation of
IRegistry.unregister
getInstance()
static getInstance(): ExpertRegistry;

Defined in: packages/nexus-agents/src/agents/experts/expert-registry.ts:79

Get the singleton instance.

Returns

ExpertRegistry

resetInstance()
static resetInstance(): void;

Defined in: packages/nexus-agents/src/agents/experts/expert-registry.ts:87

Reset the singleton instance (for testing).

Returns

void


FactoryError

Defined in: packages/nexus-agents/src/agents/experts/expert-factory.ts:37

Error specific to factory operations.

Extends

Constructors

Constructor
new FactoryError(message, options?): FactoryError;

Defined in: packages/nexus-agents/src/agents/experts/expert-factory.ts:38

Parameters
message

string

options?
cause?

Error

context?

Record<string, unknown>

Returns

FactoryError

Overrides

AgentError.constructor

Properties

cause
readonly cause: Error | undefined;

Defined in: packages/nexus-agents/src/core/errors.ts:105

Inherited from

AgentError.cause

code
readonly code: ErrorCode;

Defined in: packages/nexus-agents/src/core/errors.ts:103

Inherited from

AgentError.code

context
readonly context: Record<string, unknown> | undefined;

Defined in: packages/nexus-agents/src/core/errors.ts:104

Inherited from

AgentError.context

message
message: string;

Defined in: node_modules/.pnpm/typescript@6.0.3/node_modules/typescript/lib/lib.es5.d.ts:1075

Inherited from

AgentError.message

name
name: string;

Defined in: node_modules/.pnpm/typescript@6.0.3/node_modules/typescript/lib/lib.es5.d.ts:1074

Inherited from

AgentError.name

stack?
optional stack?: string;

Defined in: node_modules/.pnpm/typescript@6.0.3/node_modules/typescript/lib/lib.es5.d.ts:1076

Inherited from

AgentError.stack

stackTraceLimit
static stackTraceLimit: number;

Defined in: node_modules/.pnpm/@types+node@25.9.8/node_modules/@types/node/globals.d.ts:67

The Error.stackTraceLimit property specifies the number of stack frames collected by a stack trace (whether generated by new Error().stack or Error.captureStackTrace(obj)).

The default value is 10 but may be set to any valid JavaScript number. Changes will affect any stack trace captured after the value has been changed.

If set to a non-number value, or set to a negative number, stack traces will not capture any frames.

Inherited from

AgentError.stackTraceLimit

Methods

toJSON()
toJSON(): SerializedError;

Defined in: packages/nexus-agents/src/core/errors.ts:121

Serializes the error to a JSON-safe object.

Returns

SerializedError

Inherited from

AgentError.toJSON

captureStackTrace()
static captureStackTrace(targetObject, constructorOpt?): void;

Defined in: node_modules/.pnpm/@types+node@25.9.8/node_modules/@types/node/globals.d.ts:51

Creates a .stack property on targetObject, which when accessed returns a string representing the location in the code at which Error.captureStackTrace() was called.

const myObject = {};
Error.captureStackTrace(myObject);
myObject.stack;  // Similar to `new Error().stack`

The first line of the trace will be prefixed with ${myObject.name}: ${myObject.message}.

The optional constructorOpt argument accepts a function. If given, all frames above constructorOpt, including constructorOpt, will be omitted from the generated stack trace.

The constructorOpt argument is useful for hiding implementation details of error generation from the user. For instance:

function a() {
  b();
}

function b() {
  c();
}

function c() {
  // Create an error without stack trace to avoid calculating the stack trace twice.
  const { stackTraceLimit } = Error;
  Error.stackTraceLimit = 0;
  const error = new Error();
  Error.stackTraceLimit = stackTraceLimit;

  // Capture the stack trace above function b
  Error.captureStackTrace(error, b); // Neither function c, nor b is included in the stack trace
  throw error;
}

a();
Parameters
targetObject

object

constructorOpt?

Function

Returns

void

Inherited from

AgentError.captureStackTrace

prepareStackTrace()
static prepareStackTrace(err, stackTraces): any;

Defined in: node_modules/.pnpm/@types+node@25.9.8/node_modules/@types/node/globals.d.ts:55

Parameters
err

Error

stackTraces

CallSite[]

Returns

any

See

https://v8.dev/docs/stack-trace-api#customizing-stack-traces

Inherited from

AgentError.prepareStackTrace


Orchestrator

Defined in: packages/nexus-agents/src/agents/tech-lead.ts:202

Agents exports - Agent framework, Orchestrator, Experts Split from index.ts for file size compliance (Issue #285)

Extends

Constructors

Constructor
new Orchestrator(options?): Orchestrator;

Defined in: packages/nexus-agents/src/agents/tech-lead.ts:208

Parameters
options?

Partial<BaseAgentOptions> & OrchestratorExtendedOptions & { techLeadOptions?: OrchestratorOptions; } = {}

Returns

Orchestrator

Overrides

BaseAgent.constructor

Properties

adapter
protected adapter: IModelAdapter | undefined;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:98

Inherited from

BaseAgent.adapter

budgetTracker
protected readonly budgetTracker: ITokenBudgetTracker;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:97

Inherited from

BaseAgent.budgetTracker

capabilities
readonly capabilities: readonly AgentCapability[];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:95

Agent capabilities

Inherited from

BaseAgent.capabilities

config
protected config: AgentConfig | undefined;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:100

Inherited from

BaseAgent.config

emitMessageEvents
protected readonly emitMessageEvents: boolean;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:107

Inherited from

BaseAgent.emitMessageEvents

eventBus
protected readonly eventBus: ICollaborationEventBus;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:106

Inherited from

BaseAgent.eventBus

history
protected history: Message[] = [];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:102

Inherited from

BaseAgent.history

id
readonly id: string;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:93

Unique agent identifier

Inherited from

BaseAgent.id

logger
protected readonly logger: ILogger;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:99

Inherited from

BaseAgent.logger

maxTokens
protected readonly maxTokens: number;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:105

Inherited from

BaseAgent.maxTokens

role
readonly role: AgentRole;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:94

Agent role

Inherited from

BaseAgent.role

sharedState
protected sharedState: Record<string, unknown> = {};

Defined in: packages/nexus-agents/src/agents/base-agent.ts:101

Inherited from

BaseAgent.sharedState

stateMachine
protected readonly stateMachine: AgentStateMachine;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:96

Inherited from

BaseAgent.stateMachine

systemPrompt
protected readonly systemPrompt: string | undefined;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:103

Inherited from

BaseAgent.systemPrompt

temperature
protected readonly temperature: number;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:104

Inherited from

BaseAgent.temperature

Accessors

state
Get Signature
get state(): AgentState;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:176

Current state

Returns

AgentState

Current state

Inherited from

BaseAgent.state

Methods

addContextItem()
protected addContextItem(
   content, 
   priority?, 
   category?
): Promise<void>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:344

Parameters
content

string

priority?

100 | 20 | 60 | 80 | 40

category?

"system" | "task" | "active"

Returns

Promise<void>

Inherited from

BaseAgent.addContextItem

addToHistory()
protected addToHistory(message): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:331

Parameters
message

Message

Returns

void

Inherited from

BaseAgent.addToHistory

analyzeTask()
analyzeTask(task): Promise<Result<TaskAnalysis, AgentError>>;

Defined in: packages/nexus-agents/src/agents/tech-lead.ts:308

Analyze a task to understand its complexity and requirements.

Parameters
task

Task

Returns

Promise<Result<TaskAnalysis, AgentError>>

buildPrompt()
protected buildPrompt(task): Message[];

Defined in: packages/nexus-agents/src/agents/tech-lead.ts:303

Build prompt messages for task execution.

Parameters
task

Task

Returns

Message[]

Overrides

BaseAgent.buildPrompt

cleanup()
cleanup(): Promise<void>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:276

Cleanup agent resources.

Returns

Promise<void>

Inherited from

BaseAgent.cleanup

clearHistory()
protected clearHistory(): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:337

Returns

void

Inherited from

BaseAgent.clearHistory

complete()
protected complete(request): Promise<Result<CompletionResponse, AgentError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:303

Parameters
request

CompletionRequest

Returns

Promise<Result<CompletionResponse, AgentError>>

Inherited from

BaseAgent.complete

decomposeTask()
decomposeTask(task, analysis): Promise<Result<SubTask[], AgentError>>;

Defined in: packages/nexus-agents/src/agents/tech-lead.ts:350

Decompose a task into subtasks.

Parameters
task

Task

analysis

TaskAnalysis

Returns

Promise<Result<SubTask[], AgentError>>

execute()
execute(task, options?): Promise<Result<TaskResult, AgentError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:218

Execute a task.

Parameters
task

Task

Task to execute

options?

Optional execution options (#3016/#3040). signal cancels the in-flight model call when the caller’s deadline wins a race; without it, the SDK keeps running to its own 10-minute timeout after the caller has already discarded the result.

signal?

AbortSignal

Returns

Promise<Result<TaskResult, AgentError>>

Result with TaskResult or AgentError

Inherited from

BaseAgent.execute

executeTask()
protected executeTask(task): Promise<Result<TaskResult, AgentError>>;

Defined in: packages/nexus-agents/src/agents/tech-lead.ts:245

Execute a task by analyzing, decomposing (if needed), and coordinating.

Parameters
task

Task

Returns

Promise<Result<TaskResult, AgentError>>

Overrides

BaseAgent.executeTask

findResolutionForError()
protected findResolutionForError(errorMessage): ErrorResolution | undefined;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:393

Parameters
errorMessage

string

Returns

ErrorResolution | undefined

Inherited from

BaseAgent.findResolutionForError

flushMemory()
flushMemory(): Promise<Result<void, AgentMemoryError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:371

Returns

Promise<Result<void, AgentMemoryError>>

Inherited from

BaseAgent.flushMemory

getCollaborationHelper()
getCollaborationHelper(): OrchestratorCollaborationHelper;

Defined in: packages/nexus-agents/src/agents/tech-lead.ts:518

Get the collaboration helper for external use.

Returns

OrchestratorCollaborationHelper

getHistory()
protected getHistory(): Message[];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:334

Returns

Message[]

Inherited from

BaseAgent.getHistory

getMemoryState()
getMemoryState(): Readonly<AgentMemoryState> | null;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:364

Returns

Readonly<AgentMemoryState> | null

Inherited from

BaseAgent.getMemoryState

getOptions()
getOptions(): Readonly<Required<OrchestratorOptions>>;

Defined in: packages/nexus-agents/src/agents/tech-lead.ts:523

Get the Orchestrator options.

Returns

Readonly<Required<OrchestratorOptions>>

getPruningMetrics()
getPruningMetrics(): Readonly<ContextPruningMetrics>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:340

Returns

Readonly<ContextPruningMetrics>

Inherited from

BaseAgent.getPruningMetrics

getRelevantMemories()
getRelevantMemories(): readonly TypedMemoryEntry<MemoryType>[];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:367

Returns

readonly TypedMemoryEntry<MemoryType>[]

Inherited from

BaseAgent.getRelevantMemories

getTaskLearnings()
protected getTaskLearnings(taskType): readonly TaskLearning[];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:396

Parameters
taskType

string

Returns

readonly TaskLearning[]

Inherited from

BaseAgent.getTaskLearnings

handleMessage()
handleMessage(msg): Promise<Result<AgentResponse, AgentError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:261

Handle an inter-agent message and return a response.

Delivery semantics (#3222). This is a direct, awaited request/response call: the caller invokes it and holds the returned promise. It is NOT a queued or broadcast channel — that is the collaboration event bus (agents/collaboration/event-bus.ts), a fire-and-forget pub/sub with its own semantics. For this method specifically:

  • Ordering is the caller’s responsibility. Sequential awaits are handled in call order; concurrent calls carry no cross-message ordering guarantee.
  • Delivery is exactly the method invocation — there is no automatic retry or redelivery. A returned err(...) is the caller’s signal to decide whether to retry; the agent does not re-queue the message.
  • Errors surface as Result.err, not as a throw for expected conditions; the caller branches on the Result.
Parameters
msg

AgentMessage

Message to handle

Returns

Promise<Result<AgentResponse, AgentError>>

Result with AgentResponse, or AgentError on failure (not retried)

Inherited from

BaseAgent.handleMessage

hasCapability()
hasCapability(capability): boolean;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:292

Parameters
capability

AgentCapability

Returns

boolean

Inherited from

BaseAgent.hasCapability

initialize()
initialize(ctx): Promise<Result<void, AgentError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:205

Initialize the agent with context.

Parameters
ctx

AgentContext

Agent context

Returns

Promise<Result<void, AgentError>>

Result with void or AgentError

Inherited from

BaseAgent.initialize

isContextPruningEnabled()
isContextPruningEnabled(): boolean;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:358

Returns

boolean

Inherited from

BaseAgent.isContextPruningEnabled

isMemoryEnabled()
isMemoryEnabled(): boolean;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:361

Returns

boolean

Inherited from

BaseAgent.isMemoryEnabled

recordLearning()
protected recordLearning(learning): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:384

Parameters
learning

Omit<TaskLearning, "id" | "learnedAt">

Returns

void

Inherited from

BaseAgent.recordLearning

recordPattern()
protected recordPattern(p): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:387

Parameters
p

Omit<ExecutionPattern, "id" | "lastSeen" | "occurrences">

Returns

void

Inherited from

BaseAgent.recordPattern

recordResolution()
protected recordResolution(r): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:390

Parameters
r

Omit<ErrorResolution, "resolvedAt">

Returns

void

Inherited from

BaseAgent.recordResolution

selectExperts()
selectExperts(subtasks): ExpertAssignment[];

Defined in: packages/nexus-agents/src/agents/tech-lead.ts:407

Select appropriate expert agents for each subtask.

Parameters
subtasks

SubTask[]

Returns

ExpertAssignment[]

setExpertAgents()
setExpertAgents(agents): void;

Defined in: packages/nexus-agents/src/agents/tech-lead.ts:236

Set expert agents for collaboration (Issue #488). Call this to provide agents that can participate in collaborative synthesis.

Parameters
agents

Map<string, IAgent>

Returns

void

synthesizeResults()
synthesizeResults(results, originalTask?): Promise<Result<SynthesizedResult, AgentError>>;

Defined in: packages/nexus-agents/src/agents/tech-lead.ts:420

Synthesize results from multiple experts into a cohesive output.

Uses collaboration protocols for complex multi-expert synthesis (Issue #488) when enough experts and task complexity warrant it.

Parameters
results

TaskResult[]

Results to synthesize

originalTask?

Task

Optional original task for context in collaborative synthesis

Returns

Promise<Result<SynthesizedResult, AgentError>>

transformError()
protected transformError(error, taskId): AgentError;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:299

Parameters
error

unknown

taskId

string

Returns

AgentError

Inherited from

BaseAgent.transformError


ParallelProtocol

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-protocol.ts:209

Parallel collaboration protocol.

Extends

  • BaseProtocol

Constructors

Constructor
new ParallelProtocol(options?): ParallelProtocol;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-protocol.ts:42

Parameters
options?

ProtocolOptions = {}

Returns

ParallelProtocol

Inherited from
BaseProtocol.constructor

Properties

cancelled
protected cancelled: boolean = false;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-protocol.ts:40

Inherited from
BaseProtocol.cancelled
logger
protected readonly logger: ILogger;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-protocol.ts:38

Inherited from
BaseProtocol.logger
options
protected readonly options: ProtocolOptions = {};

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-protocol.ts:42

Inherited from
BaseProtocol.options
pattern
readonly pattern: "parallel";

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-protocol.ts:210

Overrides
BaseProtocol.pattern
session
protected session: CollaborationSession | null = null;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-protocol.ts:39

Inherited from
BaseProtocol.session

Methods

cancel()
cancel(reason): void;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-protocol.ts:51

Parameters
reason

string

Returns

void

Inherited from
BaseProtocol.cancel
createSession()
protected createSession(): CollaborationSession;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-protocol.ts:57

Returns

CollaborationSession

Inherited from
BaseProtocol.createSession
execute()
execute(config, agents): Promise<Result<CollaborationResult, AgentError>>;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-protocol.ts:212

Parameters
config

CollaborationConfig

agents

Map<string, IAgent>

Returns

Promise<Result<CollaborationResult, AgentError>>

Overrides
BaseProtocol.execute
executeAgentTask()
protected executeAgentTask(
   agent, 
   task, 
   previousResults?
): Promise<Result<TaskResult, AgentError>>;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-protocol.ts:62

Parameters
agent

IAgent

task

Task

previousResults?

TaskResult[]

Returns

Promise<Result<TaskResult, AgentError>>

Inherited from
BaseProtocol.executeAgentTask
validateAgents()
protected validateAgents(config, agents): Result<void, AgentError>;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-protocol.ts:85

Parameters
config

CollaborationConfig

agents

Map<string, IAgent>

Returns

Result<void, AgentError>

Inherited from
BaseProtocol.validateAgents

ProtocolFactory

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-protocol.ts:264

Factory for creating collaboration protocols.

Constructors

Constructor
new ProtocolFactory(options?): ProtocolFactory;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-protocol.ts:267

Parameters
options?

ProtocolOptions = {}

Returns

ProtocolFactory

Methods

create()
create(pattern): ICollaborationProtocol;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-protocol.ts:271

Parameters
pattern

CollaborationPattern

Returns

ICollaborationProtocol

execute()
execute(config, agents): Promise<Result<CollaborationResult, AgentError>>;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-protocol.ts:300

Parameters
config

CollaborationConfig

agents

Map<string, IAgent>

Returns

Promise<Result<CollaborationResult, AgentError>>


RecoverableExpert

Defined in: packages/nexus-agents/src/agents/experts/expert-recovery.ts:307

An Expert whose execute() applies a transient-vs-permanent recovery policy. Wrapping at execute() (not executeTask()) reuses the base state-machine auto-reset (#1060) and per-attempt heartbeat/timeout.

Extends

Constructors

Constructor
new RecoverableExpert(
   options, 
   config, 
   policy
): RecoverableExpert;

Defined in: packages/nexus-agents/src/agents/experts/expert-recovery.ts:313

Parameters
options

BaseAgentOptions

config

ExpertConfig

policy

ExpertRecoveryPolicy

Returns

RecoverableExpert

Overrides

Expert.constructor

Properties

adapter
protected adapter: IModelAdapter | undefined;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:98

Inherited from

Expert.adapter

budgetTracker
protected readonly budgetTracker: ITokenBudgetTracker;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:97

Inherited from

Expert.budgetTracker

capabilities
readonly capabilities: readonly AgentCapability[];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:95

Agent capabilities

Inherited from

Expert.capabilities

config
protected config: AgentConfig | undefined;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:100

Inherited from

Expert.config

emitMessageEvents
protected readonly emitMessageEvents: boolean;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:107

Inherited from

Expert.emitMessageEvents

eventBus
protected readonly eventBus: ICollaborationEventBus;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:106

Inherited from

Expert.eventBus

expertConfig
readonly expertConfig: ExpertConfig;

Defined in: packages/nexus-agents/src/agents/experts/expert-agent.ts:18

Inherited from

Expert.expertConfig

history
protected history: Message[] = [];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:102

Inherited from

Expert.history

id
readonly id: string;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:93

Unique agent identifier

Inherited from

Expert.id

logger
protected readonly logger: ILogger;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:99

Inherited from

Expert.logger

maxTokens
protected readonly maxTokens: number;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:105

Inherited from

Expert.maxTokens

role
readonly role: AgentRole;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:94

Agent role

Inherited from

Expert.role

sharedState
protected sharedState: Record<string, unknown> = {};

Defined in: packages/nexus-agents/src/agents/base-agent.ts:101

Inherited from

Expert.sharedState

stateMachine
protected readonly stateMachine: AgentStateMachine;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:96

Inherited from

Expert.stateMachine

systemPrompt
protected readonly systemPrompt: string | undefined;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:103

Inherited from

Expert.systemPrompt

temperature
protected readonly temperature: number;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:104

Inherited from

Expert.temperature

Accessors

metadata
Get Signature
get metadata(): Record<string, unknown> | undefined;

Defined in: packages/nexus-agents/src/agents/experts/expert-agent.ts:35

Get the expert’s metadata.

Returns

Record<string, unknown> | undefined

Inherited from

Expert.metadata

name
Get Signature
get name(): string;

Defined in: packages/nexus-agents/src/agents/experts/expert-agent.ts:28

Get the expert’s name.

Returns

string

Inherited from

Expert.name

state
Get Signature
get state(): AgentState;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:176

Current state

Returns

AgentState

Current state

Inherited from

Expert.state

Methods

addContextItem()
protected addContextItem(
   content, 
   priority?, 
   category?
): Promise<void>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:344

Parameters
content

string

priority?

100 | 20 | 60 | 80 | 40

category?

"system" | "task" | "active"

Returns

Promise<void>

Inherited from

Expert.addContextItem

addToHistory()
protected addToHistory(message): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:331

Parameters
message

Message

Returns

void

Inherited from

Expert.addToHistory

buildPrompt()
protected buildPrompt(task): Message[];

Defined in: packages/nexus-agents/src/agents/simple-agent.ts:105

Build prompt messages from a task.

Parameters
task

Task

Returns

Message[]

Inherited from

Expert.buildPrompt

cleanup()
cleanup(): Promise<void>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:276

Cleanup agent resources.

Returns

Promise<void>

Inherited from

Expert.cleanup

clearHistory()
protected clearHistory(): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:337

Returns

void

Inherited from

Expert.clearHistory

complete()
protected complete(request): Promise<Result<CompletionResponse, AgentError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:303

Parameters
request

CompletionRequest

Returns

Promise<Result<CompletionResponse, AgentError>>

Inherited from

Expert.complete

execute()
execute(task, options?): Promise<Result<TaskResult, AgentError>>;

Defined in: packages/nexus-agents/src/agents/experts/expert-recovery.ts:328

Execute a task.

Parameters
task

Task

Task to execute

options?

Optional execution options (#3016/#3040). signal cancels the in-flight model call when the caller’s deadline wins a race; without it, the SDK keeps running to its own 10-minute timeout after the caller has already discarded the result.

signal?

AbortSignal

Returns

Promise<Result<TaskResult, AgentError>>

Result with TaskResult or AgentError

Overrides

Expert.execute

executeTask()
protected executeTask(task): Promise<Result<TaskResult, AgentError>>;

Defined in: packages/nexus-agents/src/agents/simple-agent.ts:21

Execute a task by sending it to the model.

Parameters
task

Task

Returns

Promise<Result<TaskResult, AgentError>>

Inherited from

Expert.executeTask

findResolutionForError()
protected findResolutionForError(errorMessage): ErrorResolution | undefined;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:393

Parameters
errorMessage

string

Returns

ErrorResolution | undefined

Inherited from

Expert.findResolutionForError

flushMemory()
flushMemory(): Promise<Result<void, AgentMemoryError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:371

Returns

Promise<Result<void, AgentMemoryError>>

Inherited from

Expert.flushMemory

getHistory()
protected getHistory(): Message[];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:334

Returns

Message[]

Inherited from

Expert.getHistory

getMemoryState()
getMemoryState(): Readonly<AgentMemoryState> | null;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:364

Returns

Readonly<AgentMemoryState> | null

Inherited from

Expert.getMemoryState

getPruningMetrics()
getPruningMetrics(): Readonly<ContextPruningMetrics>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:340

Returns

Readonly<ContextPruningMetrics>

Inherited from

Expert.getPruningMetrics

getRelevantMemories()
getRelevantMemories(): readonly TypedMemoryEntry<MemoryType>[];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:367

Returns

readonly TypedMemoryEntry<MemoryType>[]

Inherited from

Expert.getRelevantMemories

getTaskLearnings()
protected getTaskLearnings(taskType): readonly TaskLearning[];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:396

Parameters
taskType

string

Returns

readonly TaskLearning[]

Inherited from

Expert.getTaskLearnings

handleMessage()
handleMessage(msg): Promise<Result<AgentResponse, AgentError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:261

Handle an inter-agent message and return a response.

Delivery semantics (#3222). This is a direct, awaited request/response call: the caller invokes it and holds the returned promise. It is NOT a queued or broadcast channel — that is the collaboration event bus (agents/collaboration/event-bus.ts), a fire-and-forget pub/sub with its own semantics. For this method specifically:

  • Ordering is the caller’s responsibility. Sequential awaits are handled in call order; concurrent calls carry no cross-message ordering guarantee.
  • Delivery is exactly the method invocation — there is no automatic retry or redelivery. A returned err(...) is the caller’s signal to decide whether to retry; the agent does not re-queue the message.
  • Errors surface as Result.err, not as a throw for expected conditions; the caller branches on the Result.
Parameters
msg

AgentMessage

Message to handle

Returns

Promise<Result<AgentResponse, AgentError>>

Result with AgentResponse, or AgentError on failure (not retried)

Inherited from

Expert.handleMessage

hasCapability()
hasCapability(capability): boolean;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:292

Parameters
capability

AgentCapability

Returns

boolean

Inherited from

Expert.hasCapability

initialize()
initialize(ctx): Promise<Result<void, AgentError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:205

Initialize the agent with context.

Parameters
ctx

AgentContext

Agent context

Returns

Promise<Result<void, AgentError>>

Result with void or AgentError

Inherited from

Expert.initialize

isContextPruningEnabled()
isContextPruningEnabled(): boolean;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:358

Returns

boolean

Inherited from

Expert.isContextPruningEnabled

isMemoryEnabled()
isMemoryEnabled(): boolean;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:361

Returns

boolean

Inherited from

Expert.isMemoryEnabled

recordLearning()
protected recordLearning(learning): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:384

Parameters
learning

Omit<TaskLearning, "id" | "learnedAt">

Returns

void

Inherited from

Expert.recordLearning

recordPattern()
protected recordPattern(p): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:387

Parameters
p

Omit<ExecutionPattern, "id" | "lastSeen" | "occurrences">

Returns

void

Inherited from

Expert.recordPattern

recordResolution()
protected recordResolution(r): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:390

Parameters
r

Omit<ErrorResolution, "resolvedAt">

Returns

void

Inherited from

Expert.recordResolution

transformError()
protected transformError(error, taskId): AgentError;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:299

Parameters
error

unknown

taskId

string

Returns

AgentError

Inherited from

Expert.transformError


RegistryError

Defined in: packages/nexus-agents/src/agents/experts/expert-registry.ts:18

Error specific to registry operations.

Extends

Constructors

Constructor
new RegistryError(message, options?): RegistryError;

Defined in: packages/nexus-agents/src/agents/experts/expert-registry.ts:19

Parameters
message

string

options?
cause?

Error

context?

Record<string, unknown>

Returns

RegistryError

Overrides

AgentError.constructor

Properties

cause
readonly cause: Error | undefined;

Defined in: packages/nexus-agents/src/core/errors.ts:105

Inherited from

AgentError.cause

code
readonly code: ErrorCode;

Defined in: packages/nexus-agents/src/core/errors.ts:103

Inherited from

AgentError.code

context
readonly context: Record<string, unknown> | undefined;

Defined in: packages/nexus-agents/src/core/errors.ts:104

Inherited from

AgentError.context

message
message: string;

Defined in: node_modules/.pnpm/typescript@6.0.3/node_modules/typescript/lib/lib.es5.d.ts:1075

Inherited from

AgentError.message

name
name: string;

Defined in: node_modules/.pnpm/typescript@6.0.3/node_modules/typescript/lib/lib.es5.d.ts:1074

Inherited from

AgentError.name

stack?
optional stack?: string;

Defined in: node_modules/.pnpm/typescript@6.0.3/node_modules/typescript/lib/lib.es5.d.ts:1076

Inherited from

AgentError.stack

stackTraceLimit
static stackTraceLimit: number;

Defined in: node_modules/.pnpm/@types+node@25.9.8/node_modules/@types/node/globals.d.ts:67

The Error.stackTraceLimit property specifies the number of stack frames collected by a stack trace (whether generated by new Error().stack or Error.captureStackTrace(obj)).

The default value is 10 but may be set to any valid JavaScript number. Changes will affect any stack trace captured after the value has been changed.

If set to a non-number value, or set to a negative number, stack traces will not capture any frames.

Inherited from

AgentError.stackTraceLimit

Methods

toJSON()
toJSON(): SerializedError;

Defined in: packages/nexus-agents/src/core/errors.ts:121

Serializes the error to a JSON-safe object.

Returns

SerializedError

Inherited from

AgentError.toJSON

captureStackTrace()
static captureStackTrace(targetObject, constructorOpt?): void;

Defined in: node_modules/.pnpm/@types+node@25.9.8/node_modules/@types/node/globals.d.ts:51

Creates a .stack property on targetObject, which when accessed returns a string representing the location in the code at which Error.captureStackTrace() was called.

const myObject = {};
Error.captureStackTrace(myObject);
myObject.stack;  // Similar to `new Error().stack`

The first line of the trace will be prefixed with ${myObject.name}: ${myObject.message}.

The optional constructorOpt argument accepts a function. If given, all frames above constructorOpt, including constructorOpt, will be omitted from the generated stack trace.

The constructorOpt argument is useful for hiding implementation details of error generation from the user. For instance:

function a() {
  b();
}

function b() {
  c();
}

function c() {
  // Create an error without stack trace to avoid calculating the stack trace twice.
  const { stackTraceLimit } = Error;
  Error.stackTraceLimit = 0;
  const error = new Error();
  Error.stackTraceLimit = stackTraceLimit;

  // Capture the stack trace above function b
  Error.captureStackTrace(error, b); // Neither function c, nor b is included in the stack trace
  throw error;
}

a();
Parameters
targetObject

object

constructorOpt?

Function

Returns

void

Inherited from

AgentError.captureStackTrace

prepareStackTrace()
static prepareStackTrace(err, stackTraces): any;

Defined in: node_modules/.pnpm/@types+node@25.9.8/node_modules/@types/node/globals.d.ts:55

Parameters
err

Error

stackTraces

CallSite[]

Returns

any

See

https://v8.dev/docs/stack-trace-api#customizing-stack-traces

Inherited from

AgentError.prepareStackTrace


ResultAggregator

Defined in: packages/nexus-agents/src/agents/collaboration/result-aggregator.ts:66

Agents exports - Agent framework, Orchestrator, Experts Split from index.ts for file size compliance (Issue #285)

Constructors

Constructor
new ResultAggregator(options?): ResultAggregator;

Defined in: packages/nexus-agents/src/agents/collaboration/result-aggregator.ts:72

Parameters
options?

AggregatorOptions = {}

Returns

ResultAggregator

Methods

aggregate()
aggregate(input): Result<AggregatedResult, AgentError>;

Defined in: packages/nexus-agents/src/agents/collaboration/result-aggregator.ts:82

Aggregates expert results into a final result.

Parameters
input

AggregatorInput

Returns

Result<AggregatedResult, AgentError>


ReviewProtocol

Defined in: packages/nexus-agents/src/agents/collaboration/review-protocol.ts:22

Review collaboration protocol.

Implements

Constructors

Constructor
new ReviewProtocol(options?): ReviewProtocol;

Defined in: packages/nexus-agents/src/agents/collaboration/review-protocol.ts:29

Parameters
options?

ProtocolOptions = {}

Returns

ReviewProtocol

Properties

cancelled
protected cancelled: boolean = false;

Defined in: packages/nexus-agents/src/agents/collaboration/review-protocol.ts:26

logger
protected readonly logger: ILogger;

Defined in: packages/nexus-agents/src/agents/collaboration/review-protocol.ts:24

options
protected readonly options: ProtocolOptions;

Defined in: packages/nexus-agents/src/agents/collaboration/review-protocol.ts:27

pattern
readonly pattern: "review";

Defined in: packages/nexus-agents/src/agents/collaboration/review-protocol.ts:23

Implementation of

ICollaborationProtocol.pattern

session
protected session: CollaborationSession | null = null;

Defined in: packages/nexus-agents/src/agents/collaboration/review-protocol.ts:25

Methods

cancel()
cancel(reason): void;

Defined in: packages/nexus-agents/src/agents/collaboration/review-protocol.ts:34

Parameters
reason

string

Returns

void

Implementation of

ICollaborationProtocol.cancel

execute()
execute(config, agents): Promise<Result<CollaborationResult, AgentError>>;

Defined in: packages/nexus-agents/src/agents/collaboration/review-protocol.ts:40

Parameters
config

CollaborationConfig

agents

Map<string, IAgent>

Returns

Promise<Result<CollaborationResult, AgentError>>

Implementation of

ICollaborationProtocol.execute


SecurityExpert

Defined in: packages/nexus-agents/src/agents/experts/security-expert.ts:66

SecurityExpert - Expert agent for security-related tasks.

Extends

Constructors

Constructor
new SecurityExpert(options?): SecurityExpert;

Defined in: packages/nexus-agents/src/agents/experts/security-expert.ts:69

Parameters
options?

Partial<BaseAgentOptions> & { expertOptions?: SecurityExpertOptions; } = {}

Returns

SecurityExpert

Overrides

BaseAgent.constructor

Properties

adapter
protected adapter: IModelAdapter | undefined;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:98

Inherited from

BaseAgent.adapter

budgetTracker
protected readonly budgetTracker: ITokenBudgetTracker;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:97

Inherited from

BaseAgent.budgetTracker

capabilities
readonly capabilities: readonly AgentCapability[];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:95

Agent capabilities

Inherited from

BaseAgent.capabilities

config
protected config: AgentConfig | undefined;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:100

Inherited from

BaseAgent.config

emitMessageEvents
protected readonly emitMessageEvents: boolean;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:107

Inherited from

BaseAgent.emitMessageEvents

eventBus
protected readonly eventBus: ICollaborationEventBus;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:106

Inherited from

BaseAgent.eventBus

history
protected history: Message[] = [];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:102

Inherited from

BaseAgent.history

id
readonly id: string;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:93

Unique agent identifier

Inherited from

BaseAgent.id

logger
protected readonly logger: ILogger;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:99

Inherited from

BaseAgent.logger

maxTokens
protected readonly maxTokens: number;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:105

Inherited from

BaseAgent.maxTokens

role
readonly role: AgentRole;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:94

Agent role

Inherited from

BaseAgent.role

sharedState
protected sharedState: Record<string, unknown> = {};

Defined in: packages/nexus-agents/src/agents/base-agent.ts:101

Inherited from

BaseAgent.sharedState

stateMachine
protected readonly stateMachine: AgentStateMachine;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:96

Inherited from

BaseAgent.stateMachine

systemPrompt
protected readonly systemPrompt: string | undefined;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:103

Inherited from

BaseAgent.systemPrompt

temperature
protected readonly temperature: number;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:104

Inherited from

BaseAgent.temperature

Accessors

state
Get Signature
get state(): AgentState;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:176

Current state

Returns

AgentState

Current state

Inherited from

BaseAgent.state

Methods

addContextItem()
protected addContextItem(
   content, 
   priority?, 
   category?
): Promise<void>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:344

Parameters
content

string

priority?

100 | 20 | 60 | 80 | 40

category?

"system" | "task" | "active"

Returns

Promise<void>

Inherited from

BaseAgent.addContextItem

addToHistory()
protected addToHistory(message): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:331

Parameters
message

Message

Returns

void

Inherited from

BaseAgent.addToHistory

buildPrompt()
protected buildPrompt(task): Message[];

Defined in: packages/nexus-agents/src/agents/experts/security-expert.ts:93

Parameters
task

Task

Returns

Message[]

Overrides

BaseAgent.buildPrompt

cleanup()
cleanup(): Promise<void>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:276

Cleanup agent resources.

Returns

Promise<void>

Inherited from

BaseAgent.cleanup

clearHistory()
protected clearHistory(): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:337

Returns

void

Inherited from

BaseAgent.clearHistory

complete()
protected complete(request): Promise<Result<CompletionResponse, AgentError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:303

Parameters
request

CompletionRequest

Returns

Promise<Result<CompletionResponse, AgentError>>

Inherited from

BaseAgent.complete

execute()
execute(task, options?): Promise<Result<TaskResult, AgentError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:218

Execute a task.

Parameters
task

Task

Task to execute

options?

Optional execution options (#3016/#3040). signal cancels the in-flight model call when the caller’s deadline wins a race; without it, the SDK keeps running to its own 10-minute timeout after the caller has already discarded the result.

signal?

AbortSignal

Returns

Promise<Result<TaskResult, AgentError>>

Result with TaskResult or AgentError

Inherited from

BaseAgent.execute

executeTask()
protected executeTask(task): Promise<Result<TaskResult, AgentError>>;

Defined in: packages/nexus-agents/src/agents/experts/security-expert.ts:77

Parameters
task

Task

Returns

Promise<Result<TaskResult, AgentError>>

Overrides

BaseAgent.executeTask

findResolutionForError()
protected findResolutionForError(errorMessage): ErrorResolution | undefined;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:393

Parameters
errorMessage

string

Returns

ErrorResolution | undefined

Inherited from

BaseAgent.findResolutionForError

flushMemory()
flushMemory(): Promise<Result<void, AgentMemoryError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:371

Returns

Promise<Result<void, AgentMemoryError>>

Inherited from

BaseAgent.flushMemory

getExpertOptions()
getExpertOptions(): Readonly<SecurityExpertOptions>;

Defined in: packages/nexus-agents/src/agents/experts/security-expert.ts:109

Returns

Readonly<SecurityExpertOptions>

getHistory()
protected getHistory(): Message[];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:334

Returns

Message[]

Inherited from

BaseAgent.getHistory

getMemoryState()
getMemoryState(): Readonly<AgentMemoryState> | null;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:364

Returns

Readonly<AgentMemoryState> | null

Inherited from

BaseAgent.getMemoryState

getPruningMetrics()
getPruningMetrics(): Readonly<ContextPruningMetrics>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:340

Returns

Readonly<ContextPruningMetrics>

Inherited from

BaseAgent.getPruningMetrics

getRelevantMemories()
getRelevantMemories(): readonly TypedMemoryEntry<MemoryType>[];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:367

Returns

readonly TypedMemoryEntry<MemoryType>[]

Inherited from

BaseAgent.getRelevantMemories

getTaskLearnings()
protected getTaskLearnings(taskType): readonly TaskLearning[];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:396

Parameters
taskType

string

Returns

readonly TaskLearning[]

Inherited from

BaseAgent.getTaskLearnings

handleMessage()
handleMessage(msg): Promise<Result<AgentResponse, AgentError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:261

Handle an inter-agent message and return a response.

Delivery semantics (#3222). This is a direct, awaited request/response call: the caller invokes it and holds the returned promise. It is NOT a queued or broadcast channel — that is the collaboration event bus (agents/collaboration/event-bus.ts), a fire-and-forget pub/sub with its own semantics. For this method specifically:

  • Ordering is the caller’s responsibility. Sequential awaits are handled in call order; concurrent calls carry no cross-message ordering guarantee.
  • Delivery is exactly the method invocation — there is no automatic retry or redelivery. A returned err(...) is the caller’s signal to decide whether to retry; the agent does not re-queue the message.
  • Errors surface as Result.err, not as a throw for expected conditions; the caller branches on the Result.
Parameters
msg

AgentMessage

Message to handle

Returns

Promise<Result<AgentResponse, AgentError>>

Result with AgentResponse, or AgentError on failure (not retried)

Inherited from

BaseAgent.handleMessage

hasCapability()
hasCapability(capability): boolean;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:292

Parameters
capability

AgentCapability

Returns

boolean

Inherited from

BaseAgent.hasCapability

initialize()
initialize(ctx): Promise<Result<void, AgentError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:205

Initialize the agent with context.

Parameters
ctx

AgentContext

Agent context

Returns

Promise<Result<void, AgentError>>

Result with void or AgentError

Inherited from

BaseAgent.initialize

isContextPruningEnabled()
isContextPruningEnabled(): boolean;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:358

Returns

boolean

Inherited from

BaseAgent.isContextPruningEnabled

isMemoryEnabled()
isMemoryEnabled(): boolean;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:361

Returns

boolean

Inherited from

BaseAgent.isMemoryEnabled

recordLearning()
protected recordLearning(learning): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:384

Parameters
learning

Omit<TaskLearning, "id" | "learnedAt">

Returns

void

Inherited from

BaseAgent.recordLearning

recordPattern()
protected recordPattern(p): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:387

Parameters
p

Omit<ExecutionPattern, "id" | "lastSeen" | "occurrences">

Returns

void

Inherited from

BaseAgent.recordPattern

recordResolution()
protected recordResolution(r): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:390

Parameters
r

Omit<ErrorResolution, "resolvedAt">

Returns

void

Inherited from

BaseAgent.recordResolution

transformError()
protected transformError(error, taskId): AgentError;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:299

Parameters
error

unknown

taskId

string

Returns

AgentError

Inherited from

BaseAgent.transformError


SelectionError

Defined in: packages/nexus-agents/src/agents/experts/expert-selector.ts:58

Error thrown when expert selection fails.

Extends

Constructors

Constructor
new SelectionError(message, options?): SelectionError;

Defined in: packages/nexus-agents/src/agents/experts/expert-selector.ts:59

Parameters
message

string

options?
cause?

Error

context?

Record<string, unknown>

Returns

SelectionError

Overrides

NexusError.constructor

Properties

cause
readonly cause: Error | undefined;

Defined in: packages/nexus-agents/src/core/errors.ts:105

Inherited from

NexusError.cause

code
readonly code: ErrorCode;

Defined in: packages/nexus-agents/src/core/errors.ts:103

Inherited from

NexusError.code

context
readonly context: Record<string, unknown> | undefined;

Defined in: packages/nexus-agents/src/core/errors.ts:104

Inherited from

NexusError.context

message
message: string;

Defined in: node_modules/.pnpm/typescript@6.0.3/node_modules/typescript/lib/lib.es5.d.ts:1075

Inherited from

NexusError.message

name
name: string;

Defined in: node_modules/.pnpm/typescript@6.0.3/node_modules/typescript/lib/lib.es5.d.ts:1074

Inherited from

NexusError.name

stack?
optional stack?: string;

Defined in: node_modules/.pnpm/typescript@6.0.3/node_modules/typescript/lib/lib.es5.d.ts:1076

Inherited from

NexusError.stack

stackTraceLimit
static stackTraceLimit: number;

Defined in: node_modules/.pnpm/@types+node@25.9.8/node_modules/@types/node/globals.d.ts:67

The Error.stackTraceLimit property specifies the number of stack frames collected by a stack trace (whether generated by new Error().stack or Error.captureStackTrace(obj)).

The default value is 10 but may be set to any valid JavaScript number. Changes will affect any stack trace captured after the value has been changed.

If set to a non-number value, or set to a negative number, stack traces will not capture any frames.

Inherited from

NexusError.stackTraceLimit

Methods

toJSON()
toJSON(): SerializedError;

Defined in: packages/nexus-agents/src/core/errors.ts:121

Serializes the error to a JSON-safe object.

Returns

SerializedError

Inherited from

NexusError.toJSON

captureStackTrace()
static captureStackTrace(targetObject, constructorOpt?): void;

Defined in: node_modules/.pnpm/@types+node@25.9.8/node_modules/@types/node/globals.d.ts:51

Creates a .stack property on targetObject, which when accessed returns a string representing the location in the code at which Error.captureStackTrace() was called.

const myObject = {};
Error.captureStackTrace(myObject);
myObject.stack;  // Similar to `new Error().stack`

The first line of the trace will be prefixed with ${myObject.name}: ${myObject.message}.

The optional constructorOpt argument accepts a function. If given, all frames above constructorOpt, including constructorOpt, will be omitted from the generated stack trace.

The constructorOpt argument is useful for hiding implementation details of error generation from the user. For instance:

function a() {
  b();
}

function b() {
  c();
}

function c() {
  // Create an error without stack trace to avoid calculating the stack trace twice.
  const { stackTraceLimit } = Error;
  Error.stackTraceLimit = 0;
  const error = new Error();
  Error.stackTraceLimit = stackTraceLimit;

  // Capture the stack trace above function b
  Error.captureStackTrace(error, b); // Neither function c, nor b is included in the stack trace
  throw error;
}

a();
Parameters
targetObject

object

constructorOpt?

Function

Returns

void

Inherited from

NexusError.captureStackTrace

prepareStackTrace()
static prepareStackTrace(err, stackTraces): any;

Defined in: node_modules/.pnpm/@types+node@25.9.8/node_modules/@types/node/globals.d.ts:55

Parameters
err

Error

stackTraces

CallSite[]

Returns

any

See

https://v8.dev/docs/stack-trace-api#customizing-stack-traces

Inherited from

NexusError.prepareStackTrace


SequentialProtocol

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-protocol.ts:105

Sequential collaboration protocol.

Extends

  • BaseProtocol

Constructors

Constructor
new SequentialProtocol(options?): SequentialProtocol;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-protocol.ts:42

Parameters
options?

ProtocolOptions = {}

Returns

SequentialProtocol

Inherited from
BaseProtocol.constructor

Properties

cancelled
protected cancelled: boolean = false;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-protocol.ts:40

Inherited from
BaseProtocol.cancelled
logger
protected readonly logger: ILogger;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-protocol.ts:38

Inherited from
BaseProtocol.logger
options
protected readonly options: ProtocolOptions = {};

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-protocol.ts:42

Inherited from
BaseProtocol.options
pattern
readonly pattern: "sequential";

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-protocol.ts:106

Overrides
BaseProtocol.pattern
session
protected session: CollaborationSession | null = null;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-protocol.ts:39

Inherited from
BaseProtocol.session

Methods

cancel()
cancel(reason): void;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-protocol.ts:51

Parameters
reason

string

Returns

void

Inherited from
BaseProtocol.cancel
createSession()
protected createSession(): CollaborationSession;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-protocol.ts:57

Returns

CollaborationSession

Inherited from
BaseProtocol.createSession
execute()
execute(config, agents): Promise<Result<CollaborationResult, AgentError>>;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-protocol.ts:108

Parameters
config

CollaborationConfig

agents

Map<string, IAgent>

Returns

Promise<Result<CollaborationResult, AgentError>>

Overrides
BaseProtocol.execute
executeAgentTask()
protected executeAgentTask(
   agent, 
   task, 
   previousResults?
): Promise<Result<TaskResult, AgentError>>;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-protocol.ts:62

Parameters
agent

IAgent

task

Task

previousResults?

TaskResult[]

Returns

Promise<Result<TaskResult, AgentError>>

Inherited from
BaseProtocol.executeAgentTask
validateAgents()
protected validateAgents(config, agents): Result<void, AgentError>;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-protocol.ts:85

Parameters
config

CollaborationConfig

agents

Map<string, IAgent>

Returns

Result<void, AgentError>

Inherited from
BaseProtocol.validateAgents

SimpleAgent

Defined in: packages/nexus-agents/src/agents/simple-agent.ts:17

Simple concrete agent implementation for testing and basic use cases.

This agent processes tasks by sending them directly to the model adapter and returning the response.

Extends

Extended by

Constructors

Constructor
new SimpleAgent(options): SimpleAgent;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:129

Parameters
options

BaseAgentOptions

Returns

SimpleAgent

Inherited from

BaseAgent.constructor

Properties

adapter
protected adapter: IModelAdapter | undefined;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:98

Inherited from

BaseAgent.adapter

budgetTracker
protected readonly budgetTracker: ITokenBudgetTracker;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:97

Inherited from

BaseAgent.budgetTracker

capabilities
readonly capabilities: readonly AgentCapability[];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:95

Agent capabilities

Inherited from

BaseAgent.capabilities

config
protected config: AgentConfig | undefined;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:100

Inherited from

BaseAgent.config

emitMessageEvents
protected readonly emitMessageEvents: boolean;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:107

Inherited from

BaseAgent.emitMessageEvents

eventBus
protected readonly eventBus: ICollaborationEventBus;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:106

Inherited from

BaseAgent.eventBus

history
protected history: Message[] = [];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:102

Inherited from

BaseAgent.history

id
readonly id: string;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:93

Unique agent identifier

Inherited from

BaseAgent.id

logger
protected readonly logger: ILogger;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:99

Inherited from

BaseAgent.logger

maxTokens
protected readonly maxTokens: number;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:105

Inherited from

BaseAgent.maxTokens

role
readonly role: AgentRole;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:94

Agent role

Inherited from

BaseAgent.role

sharedState
protected sharedState: Record<string, unknown> = {};

Defined in: packages/nexus-agents/src/agents/base-agent.ts:101

Inherited from

BaseAgent.sharedState

stateMachine
protected readonly stateMachine: AgentStateMachine;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:96

Inherited from

BaseAgent.stateMachine

systemPrompt
protected readonly systemPrompt: string | undefined;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:103

Inherited from

BaseAgent.systemPrompt

temperature
protected readonly temperature: number;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:104

Inherited from

BaseAgent.temperature

Accessors

state
Get Signature
get state(): AgentState;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:176

Current state

Returns

AgentState

Current state

Inherited from

BaseAgent.state

Methods

addContextItem()
protected addContextItem(
   content, 
   priority?, 
   category?
): Promise<void>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:344

Parameters
content

string

priority?

100 | 20 | 60 | 80 | 40

category?

"system" | "task" | "active"

Returns

Promise<void>

Inherited from

BaseAgent.addContextItem

addToHistory()
protected addToHistory(message): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:331

Parameters
message

Message

Returns

void

Inherited from

BaseAgent.addToHistory

buildPrompt()
protected buildPrompt(task): Message[];

Defined in: packages/nexus-agents/src/agents/simple-agent.ts:105

Build prompt messages from a task.

Parameters
task

Task

Returns

Message[]

Overrides

BaseAgent.buildPrompt

cleanup()
cleanup(): Promise<void>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:276

Cleanup agent resources.

Returns

Promise<void>

Inherited from

BaseAgent.cleanup

clearHistory()
protected clearHistory(): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:337

Returns

void

Inherited from

BaseAgent.clearHistory

complete()
protected complete(request): Promise<Result<CompletionResponse, AgentError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:303

Parameters
request

CompletionRequest

Returns

Promise<Result<CompletionResponse, AgentError>>

Inherited from

BaseAgent.complete

execute()
execute(task, options?): Promise<Result<TaskResult, AgentError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:218

Execute a task.

Parameters
task

Task

Task to execute

options?

Optional execution options (#3016/#3040). signal cancels the in-flight model call when the caller’s deadline wins a race; without it, the SDK keeps running to its own 10-minute timeout after the caller has already discarded the result.

signal?

AbortSignal

Returns

Promise<Result<TaskResult, AgentError>>

Result with TaskResult or AgentError

Inherited from

BaseAgent.execute

executeTask()
protected executeTask(task): Promise<Result<TaskResult, AgentError>>;

Defined in: packages/nexus-agents/src/agents/simple-agent.ts:21

Execute a task by sending it to the model.

Parameters
task

Task

Returns

Promise<Result<TaskResult, AgentError>>

Overrides

BaseAgent.executeTask

findResolutionForError()
protected findResolutionForError(errorMessage): ErrorResolution | undefined;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:393

Parameters
errorMessage

string

Returns

ErrorResolution | undefined

Inherited from

BaseAgent.findResolutionForError

flushMemory()
flushMemory(): Promise<Result<void, AgentMemoryError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:371

Returns

Promise<Result<void, AgentMemoryError>>

Inherited from

BaseAgent.flushMemory

getHistory()
protected getHistory(): Message[];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:334

Returns

Message[]

Inherited from

BaseAgent.getHistory

getMemoryState()
getMemoryState(): Readonly<AgentMemoryState> | null;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:364

Returns

Readonly<AgentMemoryState> | null

Inherited from

BaseAgent.getMemoryState

getPruningMetrics()
getPruningMetrics(): Readonly<ContextPruningMetrics>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:340

Returns

Readonly<ContextPruningMetrics>

Inherited from

BaseAgent.getPruningMetrics

getRelevantMemories()
getRelevantMemories(): readonly TypedMemoryEntry<MemoryType>[];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:367

Returns

readonly TypedMemoryEntry<MemoryType>[]

Inherited from

BaseAgent.getRelevantMemories

getTaskLearnings()
protected getTaskLearnings(taskType): readonly TaskLearning[];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:396

Parameters
taskType

string

Returns

readonly TaskLearning[]

Inherited from

BaseAgent.getTaskLearnings

handleMessage()
handleMessage(msg): Promise<Result<AgentResponse, AgentError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:261

Handle an inter-agent message and return a response.

Delivery semantics (#3222). This is a direct, awaited request/response call: the caller invokes it and holds the returned promise. It is NOT a queued or broadcast channel — that is the collaboration event bus (agents/collaboration/event-bus.ts), a fire-and-forget pub/sub with its own semantics. For this method specifically:

  • Ordering is the caller’s responsibility. Sequential awaits are handled in call order; concurrent calls carry no cross-message ordering guarantee.
  • Delivery is exactly the method invocation — there is no automatic retry or redelivery. A returned err(...) is the caller’s signal to decide whether to retry; the agent does not re-queue the message.
  • Errors surface as Result.err, not as a throw for expected conditions; the caller branches on the Result.
Parameters
msg

AgentMessage

Message to handle

Returns

Promise<Result<AgentResponse, AgentError>>

Result with AgentResponse, or AgentError on failure (not retried)

Inherited from

BaseAgent.handleMessage

hasCapability()
hasCapability(capability): boolean;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:292

Parameters
capability

AgentCapability

Returns

boolean

Inherited from

BaseAgent.hasCapability

initialize()
initialize(ctx): Promise<Result<void, AgentError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:205

Initialize the agent with context.

Parameters
ctx

AgentContext

Agent context

Returns

Promise<Result<void, AgentError>>

Result with void or AgentError

Inherited from

BaseAgent.initialize

isContextPruningEnabled()
isContextPruningEnabled(): boolean;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:358

Returns

boolean

Inherited from

BaseAgent.isContextPruningEnabled

isMemoryEnabled()
isMemoryEnabled(): boolean;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:361

Returns

boolean

Inherited from

BaseAgent.isMemoryEnabled

recordLearning()
protected recordLearning(learning): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:384

Parameters
learning

Omit<TaskLearning, "id" | "learnedAt">

Returns

void

Inherited from

BaseAgent.recordLearning

recordPattern()
protected recordPattern(p): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:387

Parameters
p

Omit<ExecutionPattern, "id" | "lastSeen" | "occurrences">

Returns

void

Inherited from

BaseAgent.recordPattern

recordResolution()
protected recordResolution(r): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:390

Parameters
r

Omit<ErrorResolution, "resolvedAt">

Returns

void

Inherited from

BaseAgent.recordResolution

transformError()
protected transformError(error, taskId): AgentError;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:299

Parameters
error

unknown

taskId

string

Returns

AgentError

Inherited from

BaseAgent.transformError


TestingExpert

Defined in: packages/nexus-agents/src/agents/experts/testing-expert.ts:60

TestingExpert - Expert agent for testing-related tasks.

Extends

Constructors

Constructor
new TestingExpert(options?): TestingExpert;

Defined in: packages/nexus-agents/src/agents/experts/testing-expert.ts:63

Parameters
options?

Partial<BaseAgentOptions> & { expertOptions?: TestingExpertOptions; } = {}

Returns

TestingExpert

Overrides

BaseAgent.constructor

Properties

adapter
protected adapter: IModelAdapter | undefined;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:98

Inherited from

BaseAgent.adapter

budgetTracker
protected readonly budgetTracker: ITokenBudgetTracker;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:97

Inherited from

BaseAgent.budgetTracker

capabilities
readonly capabilities: readonly AgentCapability[];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:95

Agent capabilities

Inherited from

BaseAgent.capabilities

config
protected config: AgentConfig | undefined;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:100

Inherited from

BaseAgent.config

emitMessageEvents
protected readonly emitMessageEvents: boolean;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:107

Inherited from

BaseAgent.emitMessageEvents

eventBus
protected readonly eventBus: ICollaborationEventBus;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:106

Inherited from

BaseAgent.eventBus

history
protected history: Message[] = [];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:102

Inherited from

BaseAgent.history

id
readonly id: string;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:93

Unique agent identifier

Inherited from

BaseAgent.id

logger
protected readonly logger: ILogger;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:99

Inherited from

BaseAgent.logger

maxTokens
protected readonly maxTokens: number;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:105

Inherited from

BaseAgent.maxTokens

role
readonly role: AgentRole;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:94

Agent role

Inherited from

BaseAgent.role

sharedState
protected sharedState: Record<string, unknown> = {};

Defined in: packages/nexus-agents/src/agents/base-agent.ts:101

Inherited from

BaseAgent.sharedState

stateMachine
protected readonly stateMachine: AgentStateMachine;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:96

Inherited from

BaseAgent.stateMachine

systemPrompt
protected readonly systemPrompt: string | undefined;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:103

Inherited from

BaseAgent.systemPrompt

temperature
protected readonly temperature: number;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:104

Inherited from

BaseAgent.temperature

Accessors

state
Get Signature
get state(): AgentState;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:176

Current state

Returns

AgentState

Current state

Inherited from

BaseAgent.state

Methods

addContextItem()
protected addContextItem(
   content, 
   priority?, 
   category?
): Promise<void>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:344

Parameters
content

string

priority?

100 | 20 | 60 | 80 | 40

category?

"system" | "task" | "active"

Returns

Promise<void>

Inherited from

BaseAgent.addContextItem

addToHistory()
protected addToHistory(message): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:331

Parameters
message

Message

Returns

void

Inherited from

BaseAgent.addToHistory

buildPrompt()
protected buildPrompt(task): Message[];

Defined in: packages/nexus-agents/src/agents/experts/testing-expert.ts:89

Parameters
task

Task

Returns

Message[]

Overrides

BaseAgent.buildPrompt

cleanup()
cleanup(): Promise<void>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:276

Cleanup agent resources.

Returns

Promise<void>

Inherited from

BaseAgent.cleanup

clearHistory()
protected clearHistory(): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:337

Returns

void

Inherited from

BaseAgent.clearHistory

complete()
protected complete(request): Promise<Result<CompletionResponse, AgentError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:303

Parameters
request

CompletionRequest

Returns

Promise<Result<CompletionResponse, AgentError>>

Inherited from

BaseAgent.complete

execute()
execute(task, options?): Promise<Result<TaskResult, AgentError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:218

Execute a task.

Parameters
task

Task

Task to execute

options?

Optional execution options (#3016/#3040). signal cancels the in-flight model call when the caller’s deadline wins a race; without it, the SDK keeps running to its own 10-minute timeout after the caller has already discarded the result.

signal?

AbortSignal

Returns

Promise<Result<TaskResult, AgentError>>

Result with TaskResult or AgentError

Inherited from

BaseAgent.execute

executeTask()
protected executeTask(task): Promise<Result<TaskResult, AgentError>>;

Defined in: packages/nexus-agents/src/agents/experts/testing-expert.ts:71

Parameters
task

Task

Returns

Promise<Result<TaskResult, AgentError>>

Overrides

BaseAgent.executeTask

findResolutionForError()
protected findResolutionForError(errorMessage): ErrorResolution | undefined;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:393

Parameters
errorMessage

string

Returns

ErrorResolution | undefined

Inherited from

BaseAgent.findResolutionForError

flushMemory()
flushMemory(): Promise<Result<void, AgentMemoryError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:371

Returns

Promise<Result<void, AgentMemoryError>>

Inherited from

BaseAgent.flushMemory

getExpertOptions()
getExpertOptions(): Readonly<TestingExpertOptions>;

Defined in: packages/nexus-agents/src/agents/experts/testing-expert.ts:105

Returns

Readonly<TestingExpertOptions>

getHistory()
protected getHistory(): Message[];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:334

Returns

Message[]

Inherited from

BaseAgent.getHistory

getMemoryState()
getMemoryState(): Readonly<AgentMemoryState> | null;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:364

Returns

Readonly<AgentMemoryState> | null

Inherited from

BaseAgent.getMemoryState

getPruningMetrics()
getPruningMetrics(): Readonly<ContextPruningMetrics>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:340

Returns

Readonly<ContextPruningMetrics>

Inherited from

BaseAgent.getPruningMetrics

getRelevantMemories()
getRelevantMemories(): readonly TypedMemoryEntry<MemoryType>[];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:367

Returns

readonly TypedMemoryEntry<MemoryType>[]

Inherited from

BaseAgent.getRelevantMemories

getTaskLearnings()
protected getTaskLearnings(taskType): readonly TaskLearning[];

Defined in: packages/nexus-agents/src/agents/base-agent.ts:396

Parameters
taskType

string

Returns

readonly TaskLearning[]

Inherited from

BaseAgent.getTaskLearnings

handleMessage()
handleMessage(msg): Promise<Result<AgentResponse, AgentError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:261

Handle an inter-agent message and return a response.

Delivery semantics (#3222). This is a direct, awaited request/response call: the caller invokes it and holds the returned promise. It is NOT a queued or broadcast channel — that is the collaboration event bus (agents/collaboration/event-bus.ts), a fire-and-forget pub/sub with its own semantics. For this method specifically:

  • Ordering is the caller’s responsibility. Sequential awaits are handled in call order; concurrent calls carry no cross-message ordering guarantee.
  • Delivery is exactly the method invocation — there is no automatic retry or redelivery. A returned err(...) is the caller’s signal to decide whether to retry; the agent does not re-queue the message.
  • Errors surface as Result.err, not as a throw for expected conditions; the caller branches on the Result.
Parameters
msg

AgentMessage

Message to handle

Returns

Promise<Result<AgentResponse, AgentError>>

Result with AgentResponse, or AgentError on failure (not retried)

Inherited from

BaseAgent.handleMessage

hasCapability()
hasCapability(capability): boolean;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:292

Parameters
capability

AgentCapability

Returns

boolean

Inherited from

BaseAgent.hasCapability

initialize()
initialize(ctx): Promise<Result<void, AgentError>>;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:205

Initialize the agent with context.

Parameters
ctx

AgentContext

Agent context

Returns

Promise<Result<void, AgentError>>

Result with void or AgentError

Inherited from

BaseAgent.initialize

isContextPruningEnabled()
isContextPruningEnabled(): boolean;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:358

Returns

boolean

Inherited from

BaseAgent.isContextPruningEnabled

isMemoryEnabled()
isMemoryEnabled(): boolean;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:361

Returns

boolean

Inherited from

BaseAgent.isMemoryEnabled

recordLearning()
protected recordLearning(learning): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:384

Parameters
learning

Omit<TaskLearning, "id" | "learnedAt">

Returns

void

Inherited from

BaseAgent.recordLearning

recordPattern()
protected recordPattern(p): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:387

Parameters
p

Omit<ExecutionPattern, "id" | "lastSeen" | "occurrences">

Returns

void

Inherited from

BaseAgent.recordPattern

recordResolution()
protected recordResolution(r): void;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:390

Parameters
r

Omit<ErrorResolution, "resolvedAt">

Returns

void

Inherited from

BaseAgent.recordResolution

transformError()
protected transformError(error, taskId): AgentError;

Defined in: packages/nexus-agents/src/agents/base-agent.ts:299

Parameters
error

unknown

taskId

string

Returns

AgentError

Inherited from

BaseAgent.transformError


TrinityCoordinator

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-coordinator.ts:60

Coordinates Thinker, Worker, and Verifier roles for task execution.

Constructors

Constructor
new TrinityCoordinator(options?): TrinityCoordinator;

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-coordinator.ts:67

Parameters
options?

TrinityConfig | TrinityCoordinatorOptions

Returns

TrinityCoordinator

Methods

cancel()
cancel(reason): void;

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-coordinator.ts:86

Parameters
reason

string

Returns

void

execute()
execute(options): Promise<Result<TrinityResult, AgentError>>;

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-coordinator.ts:91

Parameters
options

TrinityExecuteOptions

Returns

Promise<Result<TrinityResult, AgentError>>


WaveScheduler

Defined in: packages/nexus-agents/src/agents/wave-scheduler.ts:60

Wave scheduler for bounded parallel task execution.

Executes tasks in waves respecting concurrency limits, dependency order, output budgets, and token budgets. Each wave waits for all tasks to complete before the next wave launches.

Example

const scheduler = createWaveScheduler({ maxConcurrency: 3 });
const result = await scheduler.execute(tasks, async (task) => {
  return await runAgent(task.input);
});
console.log(`Completed in ${result.waves.length} waves`);

Constructors

Constructor
new WaveScheduler(config?, logger?): WaveScheduler;

Defined in: packages/nexus-agents/src/agents/wave-scheduler.ts:64

Parameters
config?

Partial<WaveSchedulerConfig> = {}

logger?

ILogger

Returns

WaveScheduler

Methods

buildWaves()
buildWaves<T>(tasks): WaveTask<T>[][];

Defined in: packages/nexus-agents/src/agents/wave-scheduler.ts:116

Build waves from tasks respecting dependency ordering.

Tasks with no unresolved dependencies go in the earliest possible wave. Within each wave, tasks are further split into sub-waves of maxConcurrency size.

Type Parameters
T

T

Parameters
tasks

readonly WaveTask<T>[]

Returns

WaveTask<T>[][]

execute()
execute<T>(tasks, executor): Promise<WaveExecutionResult>;

Defined in: packages/nexus-agents/src/agents/wave-scheduler.ts:72

Execute tasks in waves respecting dependencies and concurrency limits.

Type Parameters
T

T

Parameters
tasks

readonly WaveTask<T>[]

executor

WaveTaskExecutor<T>

Returns

Promise<WaveExecutionResult>

getConfig()
getConfig(): Readonly<WaveSchedulerConfig>;

Defined in: packages/nexus-agents/src/agents/wave-scheduler.ts:157

Get the scheduler configuration.

Returns

Readonly<WaveSchedulerConfig>

Interfaces

ActivationOptions

Defined in: packages/nexus-agents/src/agents/reasoning/forest-config-types.ts:190

Options for sparse activation selection.

Properties

ensureTreeCoverage
readonly ensureTreeCoverage: boolean;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-config-types.ts:198

Ensure at least one node per active tree

maxActive
readonly maxActive: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-config-types.ts:192

Maximum nodes to activate

minScore
readonly minScore: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-config-types.ts:196

Minimum score to consider for activation

strategy
readonly strategy: ActivationStrategy;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-config-types.ts:194

Strategy for selection


AgenticAdapterOptions

Defined in: packages/nexus-agents/src/agents/agentic/agentic-adapter.ts:61

Agents exports - Agent framework, Orchestrator, Experts Split from index.ts for file size compliance (Issue #285)

Properties

forceProfile?
readonly optional forceProfile?: ModelEntry;

Defined in: packages/nexus-agents/src/agents/agentic/agentic-adapter.ts:86

Force a specific behaviour profile, bypassing identity-driven lookup entirely. Reserved for tests + diagnostic runs; in production prefer modelHints so identity stays auditable.

maxConcurrent?
readonly optional maxConcurrent?: number;

Defined in: packages/nexus-agents/src/agents/agentic/agentic-adapter.ts:67

Maximum number of concurrent model API calls across all in-flight runAgent() calls. Default unlimited. Set this when the upstream provider rate-limits aggressively.

modelHints?
readonly optional modelHints?: ModelHints;

Defined in: packages/nexus-agents/src/agents/agentic/agentic-adapter.ts:74

Per-model identity overrides — gateway-renamed models, custom deployments, or anything the modelId-string parser can’t classify. Each field is optional; provided fields force, others fall through to probe / parse.

registry?
readonly optional registry?: ModelRegistry;

Defined in: packages/nexus-agents/src/agents/agentic/agentic-adapter.ts:92

Override the registry used for getEntry lookups. Defaults to the lazy global registry. Tests and multi-tenant deployments inject their own.

skipProbe?
readonly optional skipProbe?: boolean;

Defined in: packages/nexus-agents/src/agents/agentic/agentic-adapter.ts:80

Skip the IModelAdapter.listModels() probe at first runAgent. Useful when the gateway doesn’t expose /v1/models or when deterministic startup matters more than identity fidelity.


AgenticToolCall

Defined in: packages/nexus-agents/src/agents/agentic/types.ts:23

Tool call emitted by the model.

Mirrors the Anthropic Messages API tool_use ContentBlock shape; the wrapper translates whatever the underlying provider produces into this canonical form so harnesses don’t care which provider they’re talking to.

Properties

arguments
readonly arguments: Record<string, unknown>;

Defined in: packages/nexus-agents/src/agents/agentic/types.ts:29

Arguments — already JSON-parsed; provider-side is responsible for parsing.

id
readonly id: string;

Defined in: packages/nexus-agents/src/agents/agentic/types.ts:25

Unique id for this tool call, threaded back through tool_use_id.

name
readonly name: string;

Defined in: packages/nexus-agents/src/agents/agentic/types.ts:27

Tool name (must match a ToolDefinition.name from the input).


AgenticToolResult

Defined in: packages/nexus-agents/src/agents/agentic/types.ts:41

Result of a tool call, returned by the harness’s onToolCall.

content is whatever string representation of the result the model should see next turn. Convention: stringify objects, prefer one-line for primitives. isError tells the model the call failed (Anthropic surfaces this as is_error: true in the next turn’s tool_result block; other providers handle similarly).

Properties

content
readonly content: string;

Defined in: packages/nexus-agents/src/agents/agentic/types.ts:42

isError?
readonly optional isError?: boolean;

Defined in: packages/nexus-agents/src/agents/agentic/types.ts:43


AgentRunResult

Defined in: packages/nexus-agents/src/agents/agentic/types.ts:81

Final result of a successful runAgent call.

stopReason: 'agent-stopped' | 'turn-budget' | 'tool-error' | 'cancelled' is reported via the result, NOT via Result.err — partial-progress runs are gradable, and the harness inspects turns to decide.

Properties

adapterStrategy
readonly adapterStrategy: string;

Defined in: packages/nexus-agents/src/agents/agentic/types.ts:104

Strategy used to drive the loop. native:<providerId> when the underlying adapter is a known provider whose tool-use API is being threaded through; wrapper for unknown providers / custom adapters where the loop relies only on the IModelAdapter contract surface.

Eval harnesses record this so cross-provider runs are auditable.

finalContent
readonly finalContent: string;

Defined in: packages/nexus-agents/src/agents/agentic/types.ts:110

The model’s final assistant content (the response after the last tool result, when the model emits no further tool call). Empty string when the loop ended on turn-budget or cancelled.

modelId
readonly modelId: string;

Defined in: packages/nexus-agents/src/agents/agentic/types.ts:95

Model-id stamp from the underlying IModelAdapter.

providerId
readonly providerId: string;

Defined in: packages/nexus-agents/src/agents/agentic/types.ts:93

Provider-id stamp from the underlying IModelAdapter — operators read this when comparing eval results across providers, since tool-use fidelity is provider-dependent.

stopReason
readonly stopReason: AgentStopReason;

Defined in: packages/nexus-agents/src/agents/agentic/types.ts:83

totalInputTokens?
readonly optional totalInputTokens?: number;

Defined in: packages/nexus-agents/src/agents/agentic/types.ts:86

Aggregated token usage across all turns (sum of per-turn inputs/outputs).

totalOutputTokens?
readonly optional totalOutputTokens?: number;

Defined in: packages/nexus-agents/src/agents/agentic/types.ts:87

turns
readonly turns: readonly AgentTurn[];

Defined in: packages/nexus-agents/src/agents/agentic/types.ts:84

turnsUsed
readonly turnsUsed: number;

Defined in: packages/nexus-agents/src/agents/agentic/types.ts:82


AgentTurn

Defined in: packages/nexus-agents/src/agents/agentic/types.ts:50

One turn of the agent loop — model emits a tool call, harness resolves it, harness records the trace.

Properties

inputTokens?
readonly optional inputTokens?: number;

Defined in: packages/nexus-agents/src/agents/agentic/types.ts:59

Provider-reported input tokens for this turn’s API call (when available).

modelLatencyMs
readonly modelLatencyMs: number;

Defined in: packages/nexus-agents/src/agents/agentic/types.ts:55

Wall-clock time spent in the model API call that produced the tool call.

outputTokens?
readonly optional outputTokens?: number;

Defined in: packages/nexus-agents/src/agents/agentic/types.ts:61

Provider-reported output tokens for this turn’s API call (when available).

toolCall
readonly toolCall: AgenticToolCall;

Defined in: packages/nexus-agents/src/agents/agentic/types.ts:52

toolLatencyMs
readonly toolLatencyMs: number;

Defined in: packages/nexus-agents/src/agents/agentic/types.ts:57

Wall-clock time spent waiting for onToolCall to resolve.

toolResult
readonly toolResult: AgenticToolResult;

Defined in: packages/nexus-agents/src/agents/agentic/types.ts:53

turnIndex
readonly turnIndex: number;

Defined in: packages/nexus-agents/src/agents/agentic/types.ts:51


AggregatedResult

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:228

Aggregated result from multiple experts.

Properties

conflicts
conflicts: ResultConflict[];

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:232

metadata
metadata: AggregationMetadata;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:233

output
output: unknown;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:229

qualityScore
qualityScore: number;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:231

strategy
strategy: "merge" | "consensus" | "select_best" | "sequential_chain";

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:230


AggregationMetadata

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:251

Metadata about the aggregation process.

Properties

aggregatedAt
aggregatedAt: string;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:294

averageConfidence
averageConfidence: number;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:254

confidenceMeasured?
optional confidenceMeasured?: boolean;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:279

Whether averageConfidence is a measurement (#4831).

false means no contributor carried a confidence signal, so averageConfidence is a placeholder 0 and NOT a score — a session whose confidence was never assessed must not read as a confident one. The collaboration-session builder is in that position: TaskResult has no confidence field, while the ResultAggregator path computes the value from ExpertResult.confidence.

Absent means the producer predates the distinction — unknown, not measured. Mirrors ResultMetadata.tokensMeasured (#4734).

conflictCount
conflictCount: number;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:253

conflictsDetected?
optional conflictsDetected?: boolean;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:293

Whether conflicts is the outcome of a comparison (#4854).

false means nothing was compared, so an empty conflicts list and a conflictCount of 0 are the absence of a check — not the absence of disagreement. They are byte-identical to what a genuinely unanimous session produces, which is exactly why the distinction has to be stated rather than inferred. The collaboration-session builder is in that position; the ResultAggregator path compares fields pairwise.

Absent means the producer predates the distinction — unknown, not checked. Mirrors confidenceMeasured (#4831).

resultCount
resultCount: number;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:252

totalTokensUsed
totalTokensUsed: number;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:255

unmeasuredResults?
optional unmeasuredResults?: number;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:265

How many contributors reported no measured usage (#4743).

totalTokensUsed sums what was measured. A contributor whose adapter reported nothing carries a placeholder 0 (ResultMetadata.tokensMeasured === false), so without this count the total reads as complete when it is a LOWER BOUND. Absent means the aggregate predates the distinction — unknown, not zero.


AggregatorInput

Defined in: packages/nexus-agents/src/agents/collaboration/aggregator-types.ts:48

Input for aggregation.

Properties

pattern
pattern: CollaborationPattern;

Defined in: packages/nexus-agents/src/agents/collaboration/aggregator-types.ts:49

results
results: ExpertResult[];

Defined in: packages/nexus-agents/src/agents/collaboration/aggregator-types.ts:50

reviews?
optional reviews?: ReviewResponseMessage[];

Defined in: packages/nexus-agents/src/agents/collaboration/aggregator-types.ts:52

votes?
optional votes?: VoteMessage[];

Defined in: packages/nexus-agents/src/agents/collaboration/aggregator-types.ts:51


AggregatorOptions

Defined in: packages/nexus-agents/src/agents/collaboration/result-aggregator.ts:46

Options for result aggregation.

Properties

conflictResolver?
optional conflictResolver?: ConflictResolver;

Defined in: packages/nexus-agents/src/agents/collaboration/result-aggregator.ts:48

logger?
optional logger?: ILogger;

Defined in: packages/nexus-agents/src/agents/collaboration/result-aggregator.ts:47

minQualityScore?
optional minQualityScore?: number;

Defined in: packages/nexus-agents/src/agents/collaboration/result-aggregator.ts:50

qualityScorer?
optional qualityScorer?: QualityScorer;

Defined in: packages/nexus-agents/src/agents/collaboration/result-aggregator.ts:49


ApiDocumentation

Defined in: packages/nexus-agents/src/agents/experts/expert-documentation-types.ts:37

API documentation structure.

Properties

endpoints
endpoints: ApiEndpoint[];

Defined in: packages/nexus-agents/src/agents/experts/expert-documentation-types.ts:39

API endpoints or functions

types
types: ApiType[];

Defined in: packages/nexus-agents/src/agents/experts/expert-documentation-types.ts:41

Data types


ApiEndpoint

Defined in: packages/nexus-agents/src/agents/experts/expert-documentation-types.ts:47

API endpoint documentation.

Properties

description
description: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-documentation-types.ts:51

Description

example?
optional example?: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-documentation-types.ts:62

Example usage

name
name: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-documentation-types.ts:49

Endpoint name

parameters
parameters: {
  description: string;
  name: string;
  required: boolean;
  type: string;
}[];

Defined in: packages/nexus-agents/src/agents/experts/expert-documentation-types.ts:53

Parameters

description
description: string;
name
name: string;
required
required: boolean;
type
type: string;
returns
returns: {
  description: string;
  type: string;
};

Defined in: packages/nexus-agents/src/agents/experts/expert-documentation-types.ts:60

Return type

description
description: string;
type
type: string;

ApiType

Defined in: packages/nexus-agents/src/agents/experts/expert-documentation-types.ts:68

API type documentation.

Properties

description
description: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-documentation-types.ts:72

Description

name
name: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-documentation-types.ts:70

Type name

properties
properties: {
  description: string;
  name: string;
  optional: boolean;
  type: string;
}[];

Defined in: packages/nexus-agents/src/agents/experts/expert-documentation-types.ts:74

Properties

description
description: string;
name
name: string;
optional
optional: boolean;
type
type: string;

ArchitectureAnalysisResult

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:117

Architecture analysis result from ArchitectureExpert.

Extends

Properties

analysisType
analysisType: "review" | "design" | "pattern_selection";

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:119

Analysis type

components?
optional components?: SystemComponent[];

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:125

System components

confidence
confidence: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:42

Confidence score (0-1)

Inherited from

ExpertOutput.confidence

content
content: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:34

Primary result content

Inherited from

ExpertOutput.content

decisions?
optional decisions?: ArchitectureDecision[];

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:123

Design decisions

modelUsed?
optional modelUsed?: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:44

Model used for this expert’s execution (Issue #817)

Inherited from

ExpertOutput.modelUsed

patterns?
optional patterns?: ArchitecturePattern[];

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:121

Identified patterns

recommendations?
optional recommendations?: string[];

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:38

Recommendations or suggestions

Inherited from

ExpertOutput.recommendations

structuredData?
optional structuredData?: Record<string, unknown>;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:36

Structured data if applicable

Inherited from

ExpertOutput.structuredData

warnings?
optional warnings?: string[];

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:40

Warnings or issues found

Inherited from

ExpertOutput.warnings


ArchitectureDecision

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:145

Architecture decision record.

Properties

consequences
consequences: string[];

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:155

Consequences

context
context: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:151

Context

decision
decision: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:153

Decision made

id
id: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:147

Decision ID

status
status: "deprecated" | "superseded" | "accepted" | "proposed";

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:157

Status

title
title: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:149

Decision title


ArchitectureExpertOptions

Defined in: packages/nexus-agents/src/agents/experts/architecture-expert.ts:38

Configuration options for ArchitectureExpert.

Extends

Properties

additionalCapabilities?
optional additionalCapabilities?: AgentCapability[];

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:26

Custom capability extensions

Inherited from

ExpertOptions.additionalCapabilities

enableHeuristics?
optional enableHeuristics?: boolean;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:24

Enable domain-specific heuristics

Inherited from

ExpertOptions.enableHeuristics

generateADRs?
optional generateADRs?: boolean;

Defined in: packages/nexus-agents/src/agents/experts/architecture-expert.ts:42

Generate ADRs automatically

includeC4Suggestions?
optional includeC4Suggestions?: boolean;

Defined in: packages/nexus-agents/src/agents/experts/architecture-expert.ts:44

Include C4 diagram suggestions

maxTokens?
optional maxTokens?: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:22

Maximum tokens for responses

Inherited from

ExpertOptions.maxTokens

preferredStyles?
optional preferredStyles?: ArchitectureStyle[];

Defined in: packages/nexus-agents/src/agents/experts/architecture-expert.ts:40

Preferred architecture styles

qualityPriorities?
optional qualityPriorities?: QualityAttribute[];

Defined in: packages/nexus-agents/src/agents/experts/architecture-expert.ts:46

Quality attributes to prioritize

systemPromptOverride?
optional systemPromptOverride?: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:18

Custom system prompt override

Inherited from

ExpertOptions.systemPromptOverride

temperature?
optional temperature?: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:20

Temperature for completions (domain-specific default if not set)

Inherited from

ExpertOptions.temperature


ArchitecturePattern

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:131

Architecture pattern identification.

Properties

applicability
applicability: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:137

Applicability score (0-1)

category
category: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:135

Pattern category

name
name: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:133

Pattern name

tradeoffs
tradeoffs: {
  cons: string[];
  pros: string[];
};

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:139

Trade-offs

cons
cons: string[];
pros
pros: string[];

BaseAgentOptions

Defined in: packages/nexus-agents/src/agents/base-agent-types.ts:17

Options for creating a BaseAgent.

Properties

adapter?
optional adapter?: IModelAdapter;

Defined in: packages/nexus-agents/src/agents/base-agent-types.ts:25

Model adapter for LLM interactions

capabilities
capabilities: readonly AgentCapability[];

Defined in: packages/nexus-agents/src/agents/base-agent-types.ts:23

Agent capabilities

contextPruning?
optional contextPruning?: ContextPrunerAgentConfig;

Defined in: packages/nexus-agents/src/agents/base-agent-types.ts:43

Configuration for automatic context pruning (Issue #306)

emitMessageEvents?
optional emitMessageEvents?: boolean;

Defined in: packages/nexus-agents/src/agents/base-agent-types.ts:37

Whether to emit events for message handling (default: true)

eventBus?
optional eventBus?: ICollaborationEventBus;

Defined in: packages/nexus-agents/src/agents/base-agent-types.ts:35

Event bus for message observability (uses global bus if not provided)

id
id: string;

Defined in: packages/nexus-agents/src/agents/base-agent-types.ts:19

Unique agent identifier

logger?
optional logger?: ILogger;

Defined in: packages/nexus-agents/src/agents/base-agent-types.ts:27

Custom logger instance

maxTokens?
optional maxTokens?: number;

Defined in: packages/nexus-agents/src/agents/base-agent-types.ts:33

Maximum tokens for responses

memory?
optional memory?: AgentMemoryConfig;

Defined in: packages/nexus-agents/src/agents/base-agent-types.ts:45

Configuration for memory backend integration (Issue #348)

role
role: AgentRole;

Defined in: packages/nexus-agents/src/agents/base-agent-types.ts:21

Agent role

stateMachineOptions?
optional stateMachineOptions?: StateMachineOptions;

Defined in: packages/nexus-agents/src/agents/base-agent-types.ts:39

State machine options for validated state transitions

systemPrompt?
optional systemPrompt?: string;

Defined in: packages/nexus-agents/src/agents/base-agent-types.ts:29

System prompt for the agent

temperature?
optional temperature?: number;

Defined in: packages/nexus-agents/src/agents/base-agent-types.ts:31

Default temperature for completions

tokenBudget?
optional tokenBudget?: TokenBudgetConfig;

Defined in: packages/nexus-agents/src/agents/base-agent-types.ts:41

Token budget configuration for EMA-based tracking (Issue #304)


BestSolution

Defined in: packages/nexus-agents/src/agents/reasoning/forest-result-types.ts:55

The best solution found by the forest.

Properties

combinedScore
readonly combinedScore: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-result-types.ts:67

Combined score

conclusionNode
readonly conclusionNode: ReasoningNode;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-result-types.ts:61

Solution node

confidence
readonly confidence: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-result-types.ts:63

Overall confidence

path
readonly path: readonly string[];

Defined in: packages/nexus-agents/src/agents/reasoning/forest-result-types.ts:59

Path to the solution

qualityScore
readonly qualityScore: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-result-types.ts:65

Overall quality score

treeId
readonly treeId: string;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-result-types.ts:57

Tree that produced the solution


CodeAnalysisResult

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:33

Code analysis result from CodeExpert.

Extends

Properties

affectedFiles?
optional affectedFiles?: string[];

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:37

Files affected

codeChanges?
optional codeChanges?: CodeChange[];

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:39

Code changes or suggestions

confidence
confidence: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:42

Confidence score (0-1)

Inherited from

ExpertOutput.confidence

content
content: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:34

Primary result content

Inherited from

ExpertOutput.content

modelUsed?
optional modelUsed?: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:44

Model used for this expert’s execution (Issue #817)

Inherited from

ExpertOutput.modelUsed

operationType
operationType: "generation" | "debugging" | "optimization" | "refactoring";

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:35

Type of code operation performed

recommendations?
optional recommendations?: string[];

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:38

Recommendations or suggestions

Inherited from

ExpertOutput.recommendations

structuredData?
optional structuredData?: Record<string, unknown>;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:36

Structured data if applicable

Inherited from

ExpertOutput.structuredData

warnings?
optional warnings?: string[];

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:40

Warnings or issues found

Inherited from

ExpertOutput.warnings


CodeChange

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:45

Represents a single code change.

Properties

description
description: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:55

Description of change

file
file: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:47

File path

lineRange?
optional lineRange?: {
  end: number;
  start: number;
};

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:49

Line number or range

end
end: number;
start
start: number;
modified
modified: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:53

Modified code

original?
optional original?: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:51

Original code


CodeExpertOptions

Defined in: packages/nexus-agents/src/agents/experts/code-expert-helpers.ts:15

Configuration options for CodeExpert.

Extends

Properties

additionalCapabilities?
optional additionalCapabilities?: AgentCapability[];

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:26

Custom capability extensions

Inherited from

ExpertOptions.additionalCapabilities

codeStyle?
optional codeStyle?: "mixed" | "functional" | "object-oriented";

Defined in: packages/nexus-agents/src/agents/experts/code-expert-helpers.ts:19

Preferred code style (if applicable)

enableHeuristics?
optional enableHeuristics?: boolean;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:24

Enable domain-specific heuristics

Inherited from

ExpertOptions.enableHeuristics

maxTokens?
optional maxTokens?: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:22

Maximum tokens for responses

Inherited from

ExpertOptions.maxTokens

strictTypes?
optional strictTypes?: boolean;

Defined in: packages/nexus-agents/src/agents/experts/code-expert-helpers.ts:17

Enable strict type checking recommendations

systemPromptOverride?
optional systemPromptOverride?: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:18

Custom system prompt override

Inherited from

ExpertOptions.systemPromptOverride

targetLanguage?
optional targetLanguage?: string;

Defined in: packages/nexus-agents/src/agents/experts/code-expert-helpers.ts:21

Target language for code generation

temperature?
optional temperature?: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:20

Temperature for completions (domain-specific default if not set)

Inherited from

ExpertOptions.temperature


CollaborationConfig

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:67

Configuration for a collaboration session.

Properties

experts
experts: string[];

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:70

maxRetries?
optional maxRetries?: number;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:75

minVotes?
optional minVotes?: number;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:73

pattern
pattern: CollaborationPattern;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:69

requireUnanimous?
optional requireUnanimous?: boolean;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:74

sessionId
sessionId: string;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:68

task
task: Task;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:71

timeout?
optional timeout?: number;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:72


CollaborationResult

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:202

Final collaboration result.

Properties

aggregatedResult
aggregatedResult: AggregatedResult;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:205

durationMs
durationMs: number;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:207

error?
optional error?: string;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:209

expertResults
expertResults: ExpertResultSummary[];

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:206

pattern
pattern: CollaborationPattern;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:204

sessionId
sessionId: string;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:203

success
success: boolean;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:208


CollaborationSessionOptions

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-session-helpers.ts:19

Options for creating a CollaborationSession.

Properties

eventBus?
optional eventBus?: ICollaborationEventBus;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-session-helpers.ts:25

Optional event bus for cross-session event publishing

logger?
optional logger?: ILogger;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-session-helpers.ts:20

onMessage?
optional onMessage?: (message) => void;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-session-helpers.ts:22

Parameters
message

CollaborationMessage

Returns

void

onStatusChange?
optional onStatusChange?: (status) => void;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-session-helpers.ts:21

Parameters
status

SessionStatus

Returns

void

roleResolver?
optional roleResolver?: (expertId) => AgentRole;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-session-helpers.ts:23

Parameters
expertId

string

Returns

AgentRole


ComplianceStatus

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:105

Compliance check status.

Properties

findings
findings: string[];

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:111

Specific findings

framework
framework: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:107

Compliance framework

status
status: "partial" | "compliant" | "non-compliant";

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:109

Overall status


Conflict

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:160

Conflict between results.

Properties

description
description: string;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:166

Description of the conflict

resolution
resolution: string;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:168

How the conflict was resolved

subtaskId1
subtaskId1: string;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:162

First subtask ID

subtaskId2
subtaskId2: string;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:164

Second subtask ID


ContextBudget

Defined in: packages/nexus-agents/src/agents/context-manager-types.ts:32

Budget allocation for context categories. Based on PROJECT_PLAN.md recommendations.

Properties

active
active: number;

Defined in: packages/nexus-agents/src/agents/context-manager-types.ts:38

Active working content (default: 50%)

reserved
reserved: number;

Defined in: packages/nexus-agents/src/agents/context-manager-types.ts:40

Reserved for response generation (default: 15%)

system
system: number;

Defined in: packages/nexus-agents/src/agents/context-manager-types.ts:34

System instructions and project context (default: 15%)

task
task: number;

Defined in: packages/nexus-agents/src/agents/context-manager-types.ts:36

Current task description and requirements (default: 20%)


ContextItem

Defined in: packages/nexus-agents/src/agents/context-manager-types.ts:70

A piece of content in the context with its metadata.

Properties

addedAt
addedAt: number;

Defined in: packages/nexus-agents/src/agents/context-manager-types.ts:82

When this item was added

category
category: "system" | "task" | "active";

Defined in: packages/nexus-agents/src/agents/context-manager-types.ts:78

Budget category this item belongs to

content
content: string;

Defined in: packages/nexus-agents/src/agents/context-manager-types.ts:74

The content (message, text, etc.)

id
id: string;

Defined in: packages/nexus-agents/src/agents/context-manager-types.ts:72

Unique identifier for this item

metadata?
optional metadata?: Record<string, unknown>;

Defined in: packages/nexus-agents/src/agents/context-manager-types.ts:84

Optional metadata

priority
priority: ContentPriority;

Defined in: packages/nexus-agents/src/agents/context-manager-types.ts:76

Priority level for retention

tokenCount
tokenCount: number;

Defined in: packages/nexus-agents/src/agents/context-manager-types.ts:80

Token count for this item


ContextManagerConfig

Defined in: packages/nexus-agents/src/agents/context-manager-types.ts:95

Configuration for ContextManager.

Properties

adapter?
optional adapter?: IModelAdapter;

Defined in: packages/nexus-agents/src/agents/context-manager-types.ts:101

Model adapter for token counting

budget?
optional budget?: ContextBudget;

Defined in: packages/nexus-agents/src/agents/context-manager-types.ts:99

Budget allocation (defaults to DEFAULT_BUDGET)

logger?
optional logger?: ILogger;

Defined in: packages/nexus-agents/src/agents/context-manager-types.ts:103

Custom logger

maxTokens
maxTokens: number;

Defined in: packages/nexus-agents/src/agents/context-manager-types.ts:97

Maximum context window size in tokens

warningThreshold?
optional warningThreshold?: number;

Defined in: packages/nexus-agents/src/agents/context-manager-types.ts:105

Warning threshold (0-1) - warn when this % of budget is used


ContextPrunerConfig

Defined in: packages/nexus-agents/src/agents/context-pruner.ts:54

Configuration for ContextPruner.

Properties

adapter?
optional adapter?: IModelAdapter;

Defined in: packages/nexus-agents/src/agents/context-pruner.ts:56

autoTriggerThreshold?
optional autoTriggerThreshold?: number;

Defined in: packages/nexus-agents/src/agents/context-pruner.ts:61

contextManager
contextManager: ContextManager;

Defined in: packages/nexus-agents/src/agents/context-pruner.ts:55

defaultStrategy?
optional defaultStrategy?: PruningStrategy;

Defined in: packages/nexus-agents/src/agents/context-pruner.ts:58

logger?
optional logger?: ILogger;

Defined in: packages/nexus-agents/src/agents/context-pruner.ts:57

minItemsPerCategory?
optional minItemsPerCategory?: number;

Defined in: packages/nexus-agents/src/agents/context-pruner.ts:59

protectedPriority?
optional protectedPriority?: ContentPriority;

Defined in: packages/nexus-agents/src/agents/context-pruner.ts:60


ContextStats

Defined in: packages/nexus-agents/src/agents/context-manager-types.ts:120

Statistics about context usage.

Properties

availableTokens
availableTokens: number;

Defined in: packages/nexus-agents/src/agents/context-manager-types.ts:128

Available tokens (total - reserved)

categoryTokens
categoryTokens: Record<ContextItemCategory, number>;

Defined in: packages/nexus-agents/src/agents/context-manager-types.ts:124

Tokens used per category

isOverBudget
isOverBudget: boolean;

Defined in: packages/nexus-agents/src/agents/context-manager-types.ts:130

Whether any category is over budget

itemCounts
itemCounts: Record<ContextItemCategory, number>;

Defined in: packages/nexus-agents/src/agents/context-manager-types.ts:126

Number of items per category

overBudgetCategories
overBudgetCategories: ("system" | "task" | "active")[];

Defined in: packages/nexus-agents/src/agents/context-manager-types.ts:132

Categories that are over budget

totalTokens
totalTokens: number;

Defined in: packages/nexus-agents/src/agents/context-manager-types.ts:122

Total tokens currently used

usagePercentage
usagePercentage: number;

Defined in: packages/nexus-agents/src/agents/context-manager-types.ts:134

Percentage of total capacity used


CoverageMetrics

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:207

Code coverage metrics.

Properties

branch
branch: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:211

Branch coverage percentage

function
function: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:213

Function coverage percentage

line
line: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:209

Line coverage percentage

statement
statement: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:215

Statement coverage percentage

uncoveredAreas?
optional uncoveredAreas?: string[];

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:217

Uncovered areas


CreateExpertOptions

Defined in: packages/nexus-agents/src/agents/experts/expert-factory.ts:48

Options for creating an expert. (Source: Issue #476 - Wire context pruning to ExpertFactory)

Properties

adapter?
optional adapter?: IModelAdapter;

Defined in: packages/nexus-agents/src/agents/experts/expert-factory.ts:50

Model adapter to use

additionalCapabilities?
optional additionalCapabilities?: AgentCapability[];

Defined in: packages/nexus-agents/src/agents/experts/expert-factory.ts:54

Additional capabilities to add

contextPruning?
optional contextPruning?: ContextPrunerAgentConfig;

Defined in: packages/nexus-agents/src/agents/experts/expert-factory.ts:60

Context pruning configuration (Issue #476). Enables automatic memory management for long-running conversations. Since Issue #479, context pruning is enabled by default.

modelOverrides?
optional modelOverrides?: Partial<ModelPreference>;

Defined in: packages/nexus-agents/src/agents/experts/expert-factory.ts:52

Override model preferences from config

recoveryPolicy?
optional recoveryPolicy?: ExpertRecoveryPolicy;

Defined in: packages/nexus-agents/src/agents/experts/expert-factory.ts:70

Opt-in transient-vs-permanent execution recovery policy (#4286).

When present, createExpert returns a RecoverableExpert whose execute() retries transient failures (transport 429/5xx/network errors, and archetype failures classified as recoverable) with exponential backoff, and fails closed on permanent failures. When ABSENT, a plain Expert is constructed and behavior is bit-for-bit identical to the pre-#4286 path.


CreateForestInput

Defined in: packages/nexus-agents/src/agents/reasoning/forest-types.ts:167

Input for creating a new forest.

Properties

config?
readonly optional config?: Partial<ForestConfig>;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-types.ts:171

Initial configuration

initialHypotheses?
readonly optional initialHypotheses?: readonly string[];

Defined in: packages/nexus-agents/src/agents/reasoning/forest-types.ts:173

Initial tree hypotheses to explore

problem
readonly problem: string;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-types.ts:169

Problem to solve


CreateNodeInput

Defined in: packages/nexus-agents/src/agents/reasoning/forest-node-types.ts:181

Input for creating a new reasoning node.

Properties

confidence
readonly confidence: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-node-types.ts:191

Initial confidence (0-1)

content
readonly content: string;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-node-types.ts:189

Content of the reasoning step

metadata?
readonly optional metadata?: Partial<ReasoningNodeMetadata>;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-node-types.ts:193

Optional metadata

parentId
readonly parentId: string | null;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-node-types.ts:183

Parent node ID (null for root)

stepType
readonly stepType: ReasoningStepType;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-node-types.ts:187

Type of reasoning step

treeId
readonly treeId: string;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-node-types.ts:185

Tree ID this node belongs to


CreateTreeInput

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:221

Input for creating a new reasoning tree.

Properties

explorationPriority?
readonly optional explorationPriority?: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:227

Initial exploration priority

forestId
readonly forestId: string;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:223

Forest ID this tree belongs to

hypothesis
readonly hypothesis: string;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:225

Tree hypothesis or approach


CrossTreeInfo

Defined in: packages/nexus-agents/src/agents/reasoning/forest-types.ts:107

Information shared across trees for cross-pollination.

Properties

failurePatterns
readonly failurePatterns: readonly FailurePattern[];

Defined in: packages/nexus-agents/src/agents/reasoning/forest-types.ts:113

Patterns that have been proven ineffective

sharedConclusions
readonly sharedConclusions: readonly SharedConclusion[];

Defined in: packages/nexus-agents/src/agents/reasoning/forest-types.ts:109

High-confidence conclusions found in other trees

sharedInsights
readonly sharedInsights: readonly SharedInsight[];

Defined in: packages/nexus-agents/src/agents/reasoning/forest-types.ts:111

Useful intermediate results from other trees


DocumentationExpertOptions

Defined in: packages/nexus-agents/src/agents/experts/documentation-expert.ts:36

Configuration options for DocumentationExpert.

Extends

Properties

additionalCapabilities?
optional additionalCapabilities?: AgentCapability[];

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:26

Custom capability extensions

Inherited from

ExpertOptions.additionalCapabilities

audienceLevel?
optional audienceLevel?: "advanced" | "beginner" | "intermediate";

Defined in: packages/nexus-agents/src/agents/experts/documentation-expert.ts:42

Target audience level

enableHeuristics?
optional enableHeuristics?: boolean;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:24

Enable domain-specific heuristics

Inherited from

ExpertOptions.enableHeuristics

format?
optional format?: "jsdoc" | "markdown" | "tsdoc" | "rst";

Defined in: packages/nexus-agents/src/agents/experts/documentation-expert.ts:38

Documentation format

generateTOC?
optional generateTOC?: boolean;

Defined in: packages/nexus-agents/src/agents/experts/documentation-expert.ts:44

Generate table of contents

includeBadges?
optional includeBadges?: boolean;

Defined in: packages/nexus-agents/src/agents/experts/documentation-expert.ts:46

Include badges in README

includeExamples?
optional includeExamples?: boolean;

Defined in: packages/nexus-agents/src/agents/experts/documentation-expert.ts:40

Include code examples

maxTokens?
optional maxTokens?: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:22

Maximum tokens for responses

Inherited from

ExpertOptions.maxTokens

systemPromptOverride?
optional systemPromptOverride?: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:18

Custom system prompt override

Inherited from

ExpertOptions.systemPromptOverride

temperature?
optional temperature?: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:20

Temperature for completions (domain-specific default if not set)

Inherited from

ExpertOptions.temperature


DocumentationResult

Defined in: packages/nexus-agents/src/agents/experts/expert-documentation-types.ts:13

Documentation result from DocumentationExpert.

Extends

Properties

apiDocs?
optional apiDocs?: ApiDocumentation;

Defined in: packages/nexus-agents/src/agents/experts/expert-documentation-types.ts:19

API documentation

confidence
confidence: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:42

Confidence score (0-1)

Inherited from

ExpertOutput.confidence

content
content: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:34

Primary result content

Inherited from

ExpertOutput.content

documentationType
documentationType: "api" | "readme" | "guide" | "reference";

Defined in: packages/nexus-agents/src/agents/experts/expert-documentation-types.ts:15

Documentation type

modelUsed?
optional modelUsed?: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:44

Model used for this expert’s execution (Issue #817)

Inherited from

ExpertOutput.modelUsed

recommendations?
optional recommendations?: string[];

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:38

Recommendations or suggestions

Inherited from

ExpertOutput.recommendations

sections?
optional sections?: DocumentationSection[];

Defined in: packages/nexus-agents/src/agents/experts/expert-documentation-types.ts:17

Generated documentation sections

structuredData?
optional structuredData?: Record<string, unknown>;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:36

Structured data if applicable

Inherited from

ExpertOutput.structuredData

warnings?
optional warnings?: string[];

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:40

Warnings or issues found

Inherited from

ExpertOutput.warnings


DocumentationSection

Defined in: packages/nexus-agents/src/agents/experts/expert-documentation-types.ts:25

Documentation section.

Properties

content
content: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-documentation-types.ts:29

Section content

subsections?
optional subsections?: DocumentationSection[];

Defined in: packages/nexus-agents/src/agents/experts/expert-documentation-types.ts:31

Subsections

title
title: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-documentation-types.ts:27

Section title


ExecutionPlan

Defined in: packages/nexus-agents/src/agents/tech-lead.ts:140

Execution plan output structure.

The ExecutionPlan represents the Orchestrator’s analysis and decomposition of a task. It can optionally be converted to a WorkflowDefinition for replayable, static execution via the WorkflowEngine.

ExecutionPlan extends ExecutionPlanData (the pure data) with the asWorkflowDefinition conversion method.

See

ARCHITECTURE.md for the separation of concerns between Orchestrator and WorkflowEngine

Extends

  • ExecutionPlanData

Properties

analysis
analysis: TaskAnalysis;

Defined in: packages/nexus-agents/src/agents/plan-converter.ts:27

Analysis of the task complexity and requirements

Inherited from
ExecutionPlanData.analysis
assignments
assignments: ExpertAssignment[];

Defined in: packages/nexus-agents/src/agents/plan-converter.ts:31

Expert role assignments for each subtask

Inherited from
ExecutionPlanData.assignments
estimatedDuration
estimatedDuration: number;

Defined in: packages/nexus-agents/src/agents/plan-converter.ts:35

Estimated total duration in milliseconds

Inherited from
ExecutionPlanData.estimatedDuration
parallelGroups
parallelGroups: string[][];

Defined in: packages/nexus-agents/src/agents/plan-converter.ts:33

Groups of subtask IDs that can execute in parallel

Inherited from
ExecutionPlanData.parallelGroups
subtasks
subtasks: SubTask[];

Defined in: packages/nexus-agents/src/agents/plan-converter.ts:29

Decomposed subtasks (empty if task didn’t need decomposition)

Inherited from
ExecutionPlanData.subtasks
taskId
taskId: string;

Defined in: packages/nexus-agents/src/agents/plan-converter.ts:25

The original task ID this plan was created for

Inherited from
ExecutionPlanData.taskId

Methods

asWorkflowDefinition()
asWorkflowDefinition(options?): WorkflowDefinition;

Defined in: packages/nexus-agents/src/agents/tech-lead.ts:161

Convert this execution plan to a reusable WorkflowDefinition.

This “crystallizes” the dynamic plan into a static, replayable workflow that can be executed by WorkflowEngine.

Parameters
options?

PlanConversionOptions

Optional conversion configuration

Returns

WorkflowDefinition

A valid WorkflowDefinition

Example
const result = await techLead.execute(task);
const plan = result.value.output as ExecutionPlan;
const workflow = plan.asWorkflowDefinition({
  name: 'my-workflow',
  version: '1.0.0',
});
await workflowEngine.execute(workflow, inputs);

ExpertAssignment

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:96

Expert assignment for a subtask.

Properties

confidence
confidence: number;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:104

Confidence in the assignment (0-1)

expertRole
expertRole: AgentRole;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:100

Assigned expert role

ictmConfig?
optional ictmConfig?: ICTMConfig;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:106

ICTM configuration for dynamic sub-agent creation (Issue #756)

selectionReason
selectionReason: string;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:102

Reason for selection

subtaskId
subtaskId: string;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:98

Subtask ID


ExpertConfig

Defined in: packages/nexus-agents/src/agents/experts/expert-config.ts:48

Configuration for creating a dynamic expert agent.

Properties

capabilities
capabilities: AgentCapability[];

Defined in: packages/nexus-agents/src/agents/experts/expert-config.ts:58

List of capabilities this expert has

id
id: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-config.ts:50

Unique identifier for this expert

metadata?
optional metadata?: Record<string, unknown>;

Defined in: packages/nexus-agents/src/agents/experts/expert-config.ts:64

Optional metadata for extensions

modelPreference?
optional modelPreference?: ModelPreference;

Defined in: packages/nexus-agents/src/agents/experts/expert-config.ts:60

Optional model preferences

name
name: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-config.ts:52

Human-readable name

role
role: AgentRole;

Defined in: packages/nexus-agents/src/agents/experts/expert-config.ts:54

Role classification

systemPrompt
systemPrompt: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-config.ts:56

System prompt defining the expert’s behavior

toolRestrictions?
optional toolRestrictions?: {
  allowedTools?: string[];
  deniedTools?: string[];
};

Defined in: packages/nexus-agents/src/agents/experts/expert-config.ts:62

Optional tool restrictions (allowlist/denylist per role)

allowedTools?
optional allowedTools?: string[];

Tools the expert is allowed to use (allowlist — exclusive).

deniedTools?
optional deniedTools?: string[];

Tools the expert is NOT allowed to use (denylist — additive).


ExpertDefinition

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:38

Definition of an expert’s capabilities and metadata.

Properties

available
available: boolean;

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:56

Whether the expert is currently available

capabilities
capabilities: string[];

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:48

Core capabilities

description
description: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:46

Description of expert’s specialty

id
id: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:40

Unique expert identifier

name
name: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:44

Human-readable name

primaryDomain
primaryDomain: ExpertTaskDomain;

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:50

Primary domain of expertise

role
role: AgentRole;

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:42

Expert role type

secondaryDomains
secondaryDomains: ExpertTaskDomain[];

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:52

Additional domains the expert can handle

weight
weight: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:54

Base weight for scoring (0-1)


ExpertMatch

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:85

Match result for a single expert.

Properties

expertId
expertId: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:87

Expert identifier

matchedCapabilities
matchedCapabilities: string[];

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:91

Capabilities that matched the task

reasoning
reasoning: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:93

Human-readable reasoning for the match

score
score: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:89

Match score (0-1)

scoreBreakdown
scoreBreakdown: ScoreBreakdown;

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:95

Breakdown of score components


ExpertOptions

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:16

Expert-specific configuration options.

Extended by

Properties

additionalCapabilities?
optional additionalCapabilities?: AgentCapability[];

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:26

Custom capability extensions

enableHeuristics?
optional enableHeuristics?: boolean;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:24

Enable domain-specific heuristics

maxTokens?
optional maxTokens?: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:22

Maximum tokens for responses

systemPromptOverride?
optional systemPromptOverride?: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:18

Custom system prompt override

temperature?
optional temperature?: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:20

Temperature for completions (domain-specific default if not set)


ExpertOutput

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:32

Output format for expert task results.

Extended by

Properties

confidence
confidence: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:42

Confidence score (0-1)

content
content: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:34

Primary result content

modelUsed?
optional modelUsed?: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:44

Model used for this expert’s execution (Issue #817)

recommendations?
optional recommendations?: string[];

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:38

Recommendations or suggestions

structuredData?
optional structuredData?: Record<string, unknown>;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:36

Structured data if applicable

warnings?
optional warnings?: string[];

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:40

Warnings or issues found


ExpertParticipation

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:81

Expert participation record in a session.

Properties

expertId
expertId: string;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:82

joinedAt
joinedAt: string;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:84

retryCount
retryCount: number;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:87

role
role: AgentRole;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:83

status
status: "failed" | "working" | "pending" | "submitted" | "reviewing" | "voted";

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:85

submittedAt?
optional submittedAt?: string;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:86


ExpertRecoveryPolicy

Defined in: packages/nexus-agents/src/agents/experts/expert-recovery.ts:93

Opt-in recovery policy attached to an expert at creation time. All fields are optional and fall through to DEFAULT_RETRY_CONFIG (retry knobs) and this module’s detector defaults.

Properties

baseDelayMs?
optional baseDelayMs?: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-recovery.ts:101

Base backoff delay (ms). Default: DEFAULT_RETRY_CONFIG.

detectorConfig?
optional detectorConfig?: Partial<DetectorConfig>;

Defined in: packages/nexus-agents/src/agents/experts/expert-recovery.ts:107

Override the failure detector configuration.

jitterFactor?
optional jitterFactor?: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-recovery.ts:105

Jitter factor (0-1). Default: DEFAULT_RETRY_CONFIG.

maxDelayMs?
optional maxDelayMs?: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-recovery.ts:103

Maximum backoff delay (ms). Default: DEFAULT_RETRY_CONFIG.

maxRetries?
optional maxRetries?: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-recovery.ts:99

Maximum retries (attempts = maxRetries + 1). Default: EXPERT_RECOVERY_DEFAULT_MAX_RETRIES (1 → 2 attempts), NOT DEFAULT_RETRY_CONFIG.maxRetries (3).


ExpertRegisterOptions

Defined in: packages/nexus-agents/src/agents/experts/expert-registry.ts:28

Options for registering an expert.

Properties

replace?
optional replace?: boolean;

Defined in: packages/nexus-agents/src/agents/experts/expert-registry.ts:30

Whether to replace if expert with same ID exists


ExpertResult

Defined in: packages/nexus-agents/src/agents/collaboration/aggregator-types.ts:24

Expert result with metadata.

Properties

confidence?
optional confidence?: number;

Defined in: packages/nexus-agents/src/agents/collaboration/aggregator-types.ts:27

expertId
expertId: string;

Defined in: packages/nexus-agents/src/agents/collaboration/aggregator-types.ts:25

order?
optional order?: number;

Defined in: packages/nexus-agents/src/agents/collaboration/aggregator-types.ts:28

result
result: TaskResult;

Defined in: packages/nexus-agents/src/agents/collaboration/aggregator-types.ts:26


ExpertResultSummary

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:215

Summary of an expert’s contribution.

Properties

contributionScore
contributionScore: number;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:219

error?
optional error?: string;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:222

executionTimeMs
executionTimeMs: number;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:220

expertId
expertId: string;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:216

result?
optional result?: TaskResult;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:218

role
role: AgentRole;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:217

success
success: boolean;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:221


ExplorationEvent

Defined in: packages/nexus-agents/src/agents/reasoning/forest-result-types.ts:124

An event in the exploration history for debugging/analysis.

Properties

details
readonly details: Record<string, unknown>;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-result-types.ts:134

Additional details

eventType
readonly eventType: ExplorationEventType;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-result-types.ts:128

Event type

nodeId?
readonly optional nodeId?: string;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-result-types.ts:132

Node ID involved

timestamp
readonly timestamp: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-result-types.ts:126

Timestamp

treeId?
readonly optional treeId?: string;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-result-types.ts:130

Tree ID involved


FailurePattern

Defined in: packages/nexus-agents/src/agents/reasoning/forest-types.ts:86

A pattern that has been identified as ineffective.

Properties

avgFailureScore
readonly avgFailureScore: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-types.ts:92

Average quality score when this pattern appeared

occurrences
readonly occurrences: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-types.ts:90

Number of times this pattern failed

pattern
readonly pattern: string;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-types.ts:88

Pattern description


FeedbackMessage

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:153

General feedback message.

Properties

category?
optional category?: "improvement" | "concern" | "praise" | "question";

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:158

expertId
expertId: string;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:155

feedback
feedback: string;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:157

targetExpertId?
optional targetExpertId?: string;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:156

type
type: "feedback";

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:154


Forest

Defined in: packages/nexus-agents/src/agents/reasoning/forest-types.ts:136

A forest of reasoning trees with sparse activation. Coordinates multiple parallel reasoning approaches.

Properties

activationBudget
readonly activationBudget: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-types.ts:154

Maximum number of active nodes (sparse activation budget)

activeTreeIds
readonly activeTreeIds: readonly string[];

Defined in: packages/nexus-agents/src/agents/reasoning/forest-types.ts:156

Currently active tree IDs

bestPaths
readonly bestPaths: readonly PathScore[];

Defined in: packages/nexus-agents/src/agents/reasoning/forest-types.ts:147

Best paths across all trees

createdAt
readonly createdAt: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-types.ts:159

Creation timestamp

crossTreeInfo
readonly crossTreeInfo: CrossTreeInfo;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-types.ts:149

Cross-tree shared information

id
readonly id: string;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-types.ts:138

Unique forest identifier

problem
readonly problem: string;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-types.ts:140

Problem being solved

state
readonly state: ForestState;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-types.ts:145

Current state of the forest

statistics
readonly statistics: ForestStatistics;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-types.ts:151

Forest-wide statistics

trees
readonly trees: ReadonlyMap<string, ReasoningTree>;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-types.ts:142

All trees in the forest (id -> tree)

updatedAt
readonly updatedAt: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-types.ts:161

Last update timestamp


ForestConfig

Defined in: packages/nexus-agents/src/agents/reasoning/forest-config-types.ts:69

Configuration for Forest-of-Thought reasoning.

Properties

activationBudget
readonly activationBudget: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-config-types.ts:77

Total activation budget (max active nodes across forest)

activationStrategy
readonly activationStrategy: ActivationStrategy;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-config-types.ts:82

Strategy for node activation

confidenceThreshold
readonly confidenceThreshold: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-config-types.ts:93

Confidence threshold for accepting conclusions

crossTreeStrategy
readonly crossTreeStrategy: CrossTreeStrategy;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-config-types.ts:86

Strategy for cross-tree information sharing

earlyTerminationThreshold
readonly earlyTerminationThreshold: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-config-types.ts:95

Score threshold for early termination

enableCrossTreeSharing
readonly enableCrossTreeSharing: boolean;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-config-types.ts:111

Enable cross-tree information sharing

enableEarlyTermination
readonly enableEarlyTermination: boolean;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-config-types.ts:109

Enable early termination when good solution found

enableParallelExploration
readonly enableParallelExploration: boolean;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-config-types.ts:105

Enable parallel tree exploration

explorationConstant
readonly explorationConstant: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-config-types.ts:84

UCB exploration constant (for ucb strategy)

maxDepth
readonly maxDepth: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-config-types.ts:73

Maximum depth per tree

maxExplorationTimeMs
readonly maxExplorationTimeMs: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-config-types.ts:98

Maximum exploration time in ms

maxNodesPerTree
readonly maxNodesPerTree: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-config-types.ts:75

Maximum nodes per tree

maxTokensPerTree
readonly maxTokensPerTree: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-config-types.ts:102

Maximum tokens per tree

maxTrees
readonly maxTrees: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-config-types.ts:71

Maximum number of trees in the forest

minScoreThreshold
readonly minScoreThreshold: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-config-types.ts:91

Minimum score threshold for keeping nodes

nodeTimeoutMs
readonly nodeTimeoutMs: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-config-types.ts:100

Timeout per node evaluation in ms

parallelThreads
readonly parallelThreads: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-config-types.ts:107

Number of parallel exploration threads

pruningStrategy
readonly pruningStrategy: ForestPruningStrategy;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-config-types.ts:88

Strategy for pruning low-quality branches

seed
readonly seed: number | null;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-config-types.ts:116

Random seed for reproducibility (null for random)

sparsityRatio
readonly sparsityRatio: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-config-types.ts:79

Percentage of nodes to keep active (0-1)

temperature
readonly temperature: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-config-types.ts:114

Temperature for node generation (creativity vs determinism)


ForestResult

Defined in: packages/nexus-agents/src/agents/reasoning/forest-result-types.ts:155

Result of Forest-of-Thought reasoning.

Properties

bestSolution
readonly bestSolution: BestSolution | null;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-result-types.ts:162

Best solution found

conclusions
readonly conclusions: readonly ReasoningNode[];

Defined in: packages/nexus-agents/src/agents/reasoning/forest-result-types.ts:166

All conclusions reached across trees

durationMs
readonly durationMs: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-result-types.ts:176

Total duration in ms

explorationHistory?
readonly optional explorationHistory?: readonly ExplorationEvent[];

Defined in: packages/nexus-agents/src/agents/reasoning/forest-result-types.ts:181

Exploration history for analysis

finalState
readonly finalState: ForestState;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-result-types.ts:169

Final state of the forest

forestId
readonly forestId: string;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-result-types.ts:157

Forest ID

problem
readonly problem: string;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-result-types.ts:159

Original problem

statistics
readonly statistics: ForestStatistics;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-result-types.ts:173

Final statistics

terminationReason
readonly terminationReason: TerminationReason;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-result-types.ts:171

Reason for termination

topPaths
readonly topPaths: readonly PathScore[];

Defined in: packages/nexus-agents/src/agents/reasoning/forest-result-types.ts:164

All high-quality paths found

totalTokensUsed
readonly totalTokensUsed: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-result-types.ts:178

Total tokens used


ForestStatistics

Defined in: packages/nexus-agents/src/agents/reasoning/forest-state-types.ts:48

Statistics about forest exploration.

Properties

activationRatio
readonly activationRatio: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-state-types.ts:68

Activation ratio (active nodes / total nodes)

activeTrees
readonly activeTrees: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-state-types.ts:52

Number of active trees

avgTreeScore
readonly avgTreeScore: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-state-types.ts:62

Average tree score

bestPathScore
readonly bestPathScore: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-state-types.ts:60

Best path score found

maxDepth
readonly maxDepth: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-state-types.ts:58

Maximum depth across all trees

totalActiveNodes
readonly totalActiveNodes: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-state-types.ts:56

Total active nodes across all trees

totalExplorationTimeMs
readonly totalExplorationTimeMs: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-state-types.ts:66

Total exploration time in ms

totalNodes
readonly totalNodes: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-state-types.ts:54

Total nodes across all trees

totalTokensUsed
readonly totalTokensUsed: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-state-types.ts:64

Total tokens used

totalTrees
readonly totalTrees: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-state-types.ts:50

Total number of trees


GeneratedTest

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:191

Generated test case.

Properties

code
code: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:197

Test code

name
name: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:193

Test name

scenarios
scenarios: string[];

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:201

Test scenarios covered

target
target: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:199

Target function/component

type
type: "integration" | "e2e" | "unit";

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:195

Test type


IAgenticAdapter

Defined in: packages/nexus-agents/src/agents/agentic/types.ts:164

The agentic-adapter contract. Single method; all the variability lives in RunAgentArgs.

Properties

adapterStrategy
readonly adapterStrategy: string;

Defined in: packages/nexus-agents/src/agents/agentic/types.ts:167

modelId
readonly modelId: string;

Defined in: packages/nexus-agents/src/agents/agentic/types.ts:166

providerId
readonly providerId: string;

Defined in: packages/nexus-agents/src/agents/agentic/types.ts:165

Methods

runAgent()
runAgent(args): Promise<Result<AgentRunResult, AgentError>>;

Defined in: packages/nexus-agents/src/agents/agentic/types.ts:168

Parameters
args

RunAgentArgs

Returns

Promise<Result<AgentRunResult, AgentError>>


ICollaborationProtocol

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:304

Base interface for collaboration protocols.

NOTE: This interface is defined here (not in collaboration-protocol.ts) to avoid circular dependencies. Protocol implementations import this interface, and collaboration-protocol.ts imports the implementations.

Properties

pattern
readonly pattern: CollaborationPattern;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:305

Methods

cancel()
cancel(reason): void;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:310

Parameters
reason

string

Returns

void

execute()
execute(config, agents): Promise<Result<CollaborationResult, AgentError>>;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:306

Parameters
config

CollaborationConfig

agents

Map<string, IAgent>

Returns

Promise<Result<CollaborationResult, AgentError>>


ModelPreference

Defined in: packages/nexus-agents/src/agents/experts/expert-config.ts:34

Model preference configuration for an expert.

Properties

maxTokens?
optional maxTokens?: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-config.ts:42

Maximum tokens for responses

modelId?
optional modelId?: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-config.ts:38

Specific model ID

provider?
optional provider?: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-config.ts:36

Provider ID (e.g., ‘anthropic’, ‘openai’)

temperature?
optional temperature?: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-config.ts:40

Temperature for generation (0.0 - 2.0)


OrchestratorOptions

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:178

Options for the Orchestrator agent (coordination, decomposition, delegation).

Remarks

Renamed from TechLeadOptions in Issue #759. The old name is retained as a deprecated type alias.

Properties

decompositionThreshold?
optional decompositionThreshold?: number;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:182

Minimum complexity to trigger decomposition

enableParallelHints?
optional enableParallelHints?: boolean;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:184

Enable parallel execution hints

expertWeights?
optional expertWeights?: Partial<Record<AgentRole, number>>;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:186

Custom expert selection weights

maxSubtasks?
optional maxSubtasks?: number;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:180

Maximum number of subtasks to create


PathScore

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:67

A scored path through a reasoning tree from root to a target node.

Properties

breakdown
readonly breakdown: PathScoreBreakdown;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:77

Detailed score breakdown

length
readonly length: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:81

Path length (number of nodes)

path
readonly path: readonly string[];

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:71

Ordered node IDs from root to target

reachesConclusion
readonly reachesConclusion: boolean;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:79

Whether this path reaches a conclusion

score
readonly score: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:75

Overall path score (0-1)

targetNodeId
readonly targetNodeId: string;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:73

Target node (usually a conclusion)

treeId
readonly treeId: string;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:69

Tree this path belongs to


PathScoreBreakdown

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:40

Scoring breakdown for a reasoning path.

Properties

coherenceScore
readonly coherenceScore: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:46

Coherence score (logical consistency between steps)

conclusionBonus
readonly conclusionBonus: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:50

Bonus for reaching conclusion

confidenceScore
readonly confidenceScore: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:42

Average confidence across path nodes

depthFactor
readonly depthFactor: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:48

Depth penalty or bonus based on path length

qualityScore
readonly qualityScore: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:44

Average quality across path nodes


PathScoringOptions

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:100

Options for scoring a path.

Properties

coherenceWeight
readonly coherenceWeight: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:106

Weight for coherence in scoring

conclusionBonus
readonly conclusionBonus: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:110

Bonus for reaching conclusion

confidenceWeight
readonly confidenceWeight: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:102

Weight for confidence in scoring

depthPenalty
readonly depthPenalty: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:108

Penalty per depth level

qualityWeight
readonly qualityWeight: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:104

Weight for quality in scoring


ProtocolOptions

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:320

Options for protocol execution.

NOTE: sessionOptions uses a generic Record type to avoid circular dependency with collaboration-session.ts. The actual type is CollaborationSessionOptions which is defined in collaboration-session.ts.

Properties

continueOnFailure?
optional continueOnFailure?: boolean;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:325

logger?
optional logger?: ILogger;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:321

sequentialDelay?
optional sequentialDelay?: number;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:324

sessionOptions?
optional sessionOptions?: Record<string, unknown>;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:323

Session options - accepts CollaborationSessionOptions from collaboration-session.ts


PruneOptions

Defined in: packages/nexus-agents/src/agents/context-pruner.ts:82

Options for a pruning operation.

Properties

categories?
optional categories?: ("system" | "task" | "active")[];

Defined in: packages/nexus-agents/src/agents/context-pruner.ts:85

hierarchicalOptions?
optional hierarchicalOptions?: Partial<HierarchicalOptions>;

Defined in: packages/nexus-agents/src/agents/context-pruner.ts:88

semanticOptions?
optional semanticOptions?: Partial<SemanticOptions>;

Defined in: packages/nexus-agents/src/agents/context-pruner.ts:89

slidingWindowOptions?
optional slidingWindowOptions?: Partial<SlidingWindowOptions>;

Defined in: packages/nexus-agents/src/agents/context-pruner.ts:87

strategy?
optional strategy?: PruningStrategy;

Defined in: packages/nexus-agents/src/agents/context-pruner.ts:84

summarizationPrompt?
optional summarizationPrompt?: string;

Defined in: packages/nexus-agents/src/agents/context-pruner.ts:86

targetTokens?
optional targetTokens?: number;

Defined in: packages/nexus-agents/src/agents/context-pruner.ts:83


PruneResult

Defined in: packages/nexus-agents/src/agents/pruning-strategies-types.ts:240

Result of a pruning operation.

Properties

removedItems
removedItems: ContextItem[];

Defined in: packages/nexus-agents/src/agents/pruning-strategies-types.ts:241

summarizedItems
summarizedItems: ContextItem[];

Defined in: packages/nexus-agents/src/agents/pruning-strategies-types.ts:242

summaryItem?
optional summaryItem?: ContextItem;

Defined in: packages/nexus-agents/src/agents/pruning-strategies-types.ts:243

targetReached
targetReached: boolean;

Defined in: packages/nexus-agents/src/agents/pruning-strategies-types.ts:245

tokensFreed
tokensFreed: number;

Defined in: packages/nexus-agents/src/agents/pruning-strategies-types.ts:244


QueryOptions

Defined in: packages/nexus-agents/src/agents/experts/expert-registry.ts:36

Query options for finding experts.

Properties

anyCapability?
optional anyCapability?: AgentCapability[];

Defined in: packages/nexus-agents/src/agents/experts/expert-registry.ts:42

Filter by capability (expert must have at least one)

capabilities?
optional capabilities?: AgentCapability[];

Defined in: packages/nexus-agents/src/agents/experts/expert-registry.ts:40

Filter by capability (expert must have all specified)

limit?
optional limit?: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-registry.ts:44

Maximum number of results

role?
optional role?: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-registry.ts:38

Filter by role


ReasoningNode

Defined in: packages/nexus-agents/src/agents/reasoning/forest-node-types.ts:117

A single reasoning step in a reasoning tree. Represents an atomic unit of thought with content, scoring, and metadata.

Properties

activationScore
readonly activationScore: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-node-types.ts:141

Activation score determining priority (higher = more likely to activate)

children
readonly children: readonly string[];

Defined in: packages/nexus-agents/src/agents/reasoning/forest-node-types.ts:125

Child node IDs

confidence
readonly confidence: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-node-types.ts:144

Confidence in this reasoning step (0-1)

content
readonly content: string;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-node-types.ts:132

The reasoning content/thought at this step

createdAt
readonly createdAt: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-node-types.ts:151

Creation timestamp

depth
readonly depth: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-node-types.ts:127

Depth in the tree (0 for root)

estimatedValue
readonly estimatedValue: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-node-types.ts:148

Estimated value for path selection (like MCTS value)

id
readonly id: string;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-node-types.ts:119

Unique node identifier

isActive
readonly isActive: boolean;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-node-types.ts:139

Whether this node is currently activated (for sparse activation)

metadata
readonly metadata: ReasoningNodeMetadata;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-node-types.ts:134

Optional structured data associated with this step

parentId
readonly parentId: string | null;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-node-types.ts:123

Parent node ID (null for root nodes)

qualityScore
readonly qualityScore: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-node-types.ts:146

Quality score from evaluation (0-1)

state
readonly state: NodeState;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-node-types.ts:137

Current state of this node

stepType
readonly stepType: ReasoningStepType;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-node-types.ts:130

Type of reasoning step

treeId
readonly treeId: string;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-node-types.ts:121

ID of the tree this node belongs to

updatedAt
readonly updatedAt: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-node-types.ts:153

Last update timestamp


ReasoningNodeMetadata

Defined in: packages/nexus-agents/src/agents/reasoning/forest-node-types.ts:89

Metadata associated with a reasoning node.

Properties

crossReferences?
readonly optional crossReferences?: readonly string[];

Defined in: packages/nexus-agents/src/agents/reasoning/forest-node-types.ts:97

References to other nodes that informed this reasoning

custom?
readonly optional custom?: Record<string, unknown>;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-node-types.ts:99

Custom key-value pairs for extensibility

generationTimeMs?
readonly optional generationTimeMs?: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-node-types.ts:95

Time taken to generate this node in ms

source?
readonly optional source?: string;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-node-types.ts:91

Source of this reasoning (model, tool, etc.)

tokensUsed?
readonly optional tokensUsed?: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-node-types.ts:93

Tokens used to generate this node


ReasoningTree

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:172

A reasoning tree containing nodes organized hierarchically. Each tree explores one approach to solving the problem.

Properties

bestPaths
readonly bestPaths: readonly PathScore[];

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:192

Best path(s) found in this tree

createdAt
readonly createdAt: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:197

Creation timestamp

explorationPriority
readonly explorationPriority: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:187

Priority for exploration (higher = explore first)

forestId
readonly forestId: string;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:176

ID of the forest this tree belongs to

hypothesis
readonly hypothesis: string;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:190

Tree hypothesis or approach description

id
readonly id: string;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:174

Unique tree identifier

nodes
readonly nodes: ReadonlyMap<string, ReasoningNode>;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:180

All nodes in this tree (id -> node)

overallScore
readonly overallScore: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:185

Overall tree score for ranking (0-1)

rootId
readonly rootId: string;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:178

Root node ID

state
readonly state: TreeState;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:183

Current state of the tree

statistics
readonly statistics: TreeStatistics;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:194

Tree statistics

updatedAt
readonly updatedAt: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:199

Last update timestamp


RegistryStats

Defined in: packages/nexus-agents/src/agents/experts/expert-registry.ts:51

Statistics about the registry. Extends IRegistryStats for interface compatibility (ADR-0012).

Extends

  • IRegistryStats

Indexable

[key: string]: unknown

Additional stats specific to the registry type

Properties

byCapability
byCapability: Record<string, number>;

Defined in: packages/nexus-agents/src/agents/experts/expert-registry.ts:57

Count by capability

byRole
byRole: Record<string, number>;

Defined in: packages/nexus-agents/src/agents/experts/expert-registry.ts:55

Count by role

total
total: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-registry.ts:53

Total number of registered experts (IRegistryStats alias)

Overrides
IRegistryStats.total

ResultConflict

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:239

Conflict between expert results.

Properties

description
description: string;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:243

expert1Id
expert1Id: string;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:240

expert2Id
expert2Id: string;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:241

field
field: string;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:242

resolution
resolution: "unresolved" | "merged" | "expert1" | "expert2";

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:244

resolutionReason?
optional resolutionReason?: string;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:245


ResultSubmissionMessage

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:117

Result submission from an expert.

Properties

confidence?
optional confidence?: number;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:121

expertId
expertId: string;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:119

notes?
optional notes?: string;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:122

result
result: TaskResult;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:120

type
type: "result_submission";

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:118


ResultSummary

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:146

Summary of a single result.

Properties

contributions
contributions: string[];

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:154

Key contributions to final output

quality
quality: number;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:152

Quality of this result (0-1)

subtaskId
subtaskId: string;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:148

Subtask ID

summary
summary: string;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:150

Brief summary of the output


ReviewRequestMessage

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:128

Review request from one expert to another.

Properties

artifact
artifact: unknown;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:132

criteria?
optional criteria?: string[];

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:133

deadline?
optional deadline?: string;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:134

fromExpert
fromExpert: string;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:130

toExpert
toExpert: string;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:131

type
type: "review_request";

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:129


ReviewResponseMessage

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:140

Review response from a reviewer.

Properties

approved
approved: boolean;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:144

feedback
feedback: string;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:145

requesterId
requesterId: string;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:143

reviewerId
reviewerId: string;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:142

severity?
optional severity?: "critical" | "none" | "major" | "minor";

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:147

suggestions?
optional suggestions?: string[];

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:146

type
type: "review_response";

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:141


RunAgentArgs

Defined in: packages/nexus-agents/src/agents/agentic/types.ts:141

Arguments to runAgent.

onToolCall is the harness’s tool-router; the adapter awaits its Promise<ToolResult> so synchronous and async harness execution both work. Per-tool timeouts are the harness’s responsibility (the adapter doesn’t impose one — see #2529 design notes).

onTurn (optional) fires once after each turn completes, giving operators incremental progress visibility.

signal (optional) propagates external cancellation as stopReason: 'cancelled'.

Properties

maxTokens?
readonly optional maxTokens?: number;

Defined in: packages/nexus-agents/src/agents/agentic/types.ts:157

Per-turn maxTokens passed through to IModelAdapter.complete.

onToolCall
readonly onToolCall: (call) => Promise<AgenticToolResult>;

Defined in: packages/nexus-agents/src/agents/agentic/types.ts:151

Parameters
call

AgenticToolCall

Returns

Promise<AgenticToolResult>

onTurn?
readonly optional onTurn?: (turn) => void;

Defined in: packages/nexus-agents/src/agents/agentic/types.ts:152

Parameters
turn

AgentTurn

Returns

void

signal?
readonly optional signal?: AbortSignal;

Defined in: packages/nexus-agents/src/agents/agentic/types.ts:153

systemPrompt
readonly systemPrompt: string;

Defined in: packages/nexus-agents/src/agents/agentic/types.ts:142

temperature?
readonly optional temperature?: number;

Defined in: packages/nexus-agents/src/agents/agentic/types.ts:155

Sampling temperature passed through to IModelAdapter.complete.

tools
readonly tools: readonly ToolDefinition[];

Defined in: packages/nexus-agents/src/agents/agentic/types.ts:144

turnBudget?
readonly optional turnBudget?: number;

Defined in: packages/nexus-agents/src/agents/agentic/types.ts:150

Maximum agent turns. When omitted, the adapter uses the resolved model’s profile.maxRecommendedTurnBudget (claude-opus = 20, o-reasoning = 25, claude-haiku / gemini-flash = 8, defaults to 10).

userPrompt
readonly userPrompt: string;

Defined in: packages/nexus-agents/src/agents/agentic/types.ts:143


ScoreBreakdown

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:71

Breakdown of how the match score was calculated.

Properties

capabilityScore
capabilityScore: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:73

Score from capability matching (0-1)

domainScore
domainScore: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:75

Score from domain alignment (0-1)

finalScore
finalScore: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:79

Combined final score (0-1)

weightScore
weightScore: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:77

Score from weight adjustment (0-1)


SecurityAnalysisResult

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:61

Security analysis result from SecurityExpert.

Extends

Properties

compliance?
optional compliance?: ComplianceStatus;

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:79

Compliance status

confidence
confidence: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:42

Confidence score (0-1)

Inherited from

ExpertOutput.confidence

content
content: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:34

Primary result content

Inherited from

ExpertOutput.content

findingsCoverage?
optional findingsCoverage?: "unmeasured" | "partial" | "complete";

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:77

Coverage of the model-supplied findings after validation. When findingsCoverage === 'unmeasured', securityScore is 0: a fail-closed placeholder and the worst score, so a reader that ignores the marker errs toward blocking, never toward a clean 100. partial keeps the score over the validated findings; complete is unchanged.

modelUsed?
optional modelUsed?: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:44

Model used for this expert’s execution (Issue #817)

Inherited from

ExpertOutput.modelUsed

recommendations?
optional recommendations?: string[];

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:38

Recommendations or suggestions

Inherited from

ExpertOutput.recommendations

securityScore
securityScore: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:70

Security score (0-100). When findingsCoverage === 'unmeasured', securityScore is 0: a fail-closed placeholder and the worst score, so a reader that ignores the marker errs toward blocking, never toward a clean 100. partial scores only the validated findings; complete is unchanged.

structuredData?
optional structuredData?: Record<string, unknown>;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:36

Structured data if applicable

Inherited from

ExpertOutput.structuredData

vulnerabilities
vulnerabilities: Vulnerability[];

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:63

Vulnerabilities found

warnings?
optional warnings?: string[];

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:40

Warnings or issues found

Inherited from

ExpertOutput.warnings


SecurityExpertOptions

Defined in: packages/nexus-agents/src/agents/experts/security-expert.ts:40

Configuration options for SecurityExpert.

Extends

Properties

additionalCapabilities?
optional additionalCapabilities?: AgentCapability[];

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:26

Custom capability extensions

Inherited from

ExpertOptions.additionalCapabilities

complianceFrameworks?
optional complianceFrameworks?: string[];

Defined in: packages/nexus-agents/src/agents/experts/security-expert.ts:42

Compliance frameworks to check

enableCweMapping?
optional enableCweMapping?: boolean;

Defined in: packages/nexus-agents/src/agents/experts/security-expert.ts:46

Enable detailed CWE mappings

enableHeuristics?
optional enableHeuristics?: boolean;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:24

Enable domain-specific heuristics

Inherited from

ExpertOptions.enableHeuristics

focusAreas?
optional focusAreas?: SecurityFocusArea[];

Defined in: packages/nexus-agents/src/agents/experts/security-expert.ts:48

Security focus areas

maxTokens?
optional maxTokens?: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:22

Maximum tokens for responses

Inherited from

ExpertOptions.maxTokens

minSeverity?
optional minSeverity?: "critical" | "info" | "high" | "low" | "medium";

Defined in: packages/nexus-agents/src/agents/experts/security-expert.ts:44

Minimum severity to report

systemPromptOverride?
optional systemPromptOverride?: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:18

Custom system prompt override

Inherited from

ExpertOptions.systemPromptOverride

temperature?
optional temperature?: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:20

Temperature for completions (domain-specific default if not set)

Inherited from

ExpertOptions.temperature


SelectionExpertRegistry

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:60

Registry of available experts.

Methods

getAll()
getAll(): ExpertDefinition[];

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:61

Returns

ExpertDefinition[]

getAvailable()
getAvailable(): ExpertDefinition[];

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:65

Returns

ExpertDefinition[]

getByDomain()
getByDomain(domain): ExpertDefinition[];

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:64

Parameters
domain

ExpertTaskDomain

Returns

ExpertDefinition[]

getById()
getById(id): ExpertDefinition | undefined;

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:62

Parameters
id

string

Returns

ExpertDefinition | undefined

getByRole()
getByRole(role): ExpertDefinition[];

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:63

Parameters
role

AgentRole

Returns

ExpertDefinition[]


SelectionOptions

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:108

Options for expert selection.

Properties

capabilityWeights?
optional capabilityWeights?: Record<string, number>;

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:111

excludeExperts?
optional excludeExperts?: string[];

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:113

forceCollaboration?
optional forceCollaboration?: boolean;

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:114

maxAlternatives?
optional maxAlternatives?: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:110

minScore?
optional minScore?: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:109

preferredDomains?
optional preferredDomains?: ExpertTaskDomain[];

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:112


SelectionResult

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:99

Result of expert selection.

Properties

alternatives
alternatives: ExpertMatch[];

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:101

confidence
confidence: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:104

primary
primary: ExpertMatch;

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:100

requiresCollaboration
requiresCollaboration: boolean;

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:102

suggestedPattern?
optional suggestedPattern?: ExpertCollaborationPatternType;

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:103


SessionState

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:186

Aggregated session status.

Properties

completedAt?
optional completedAt?: string;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:194

config
config: CollaborationConfig;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:187

error?
optional error?: string;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:195

messageLog
messageLog: CollaborationMessage[];

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:196

participants
participants: ExpertParticipation[];

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:189

results
results: Map<string, TaskResult>;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:190

reviews
reviews: ReviewResponseMessage[];

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:191

startedAt
startedAt: string;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:193

status
status: SessionStatus;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:188

votes
votes: VoteMessage[];

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:192


SharedConclusion

Defined in: packages/nexus-agents/src/agents/reasoning/forest-types.ts:35

A conclusion shared across trees for cross-pollination.

Properties

confidence
readonly confidence: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-types.ts:43

Confidence in this conclusion

content
readonly content: string;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-types.ts:41

The conclusion content

qualityScore
readonly qualityScore: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-types.ts:45

Quality score

sourceNodeId
readonly sourceNodeId: string;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-types.ts:39

Source node ID

sourceTreeId
readonly sourceTreeId: string;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-types.ts:37

Source tree ID


SharedInsight

Defined in: packages/nexus-agents/src/agents/reasoning/forest-types.ts:62

An insight shared across trees.

Properties

content
readonly content: string;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-types.ts:68

The insight content

relevance
readonly relevance: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-types.ts:70

Relevance score for current exploration

sourceNodeId
readonly sourceNodeId: string;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-types.ts:66

Source node ID

sourceTreeId
readonly sourceTreeId: string;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-types.ts:64

Source tree ID


StateMachineOptions

Defined in: packages/nexus-agents/src/agents/state-machine-types.ts:57

State machine options.

Properties

initialState?
optional initialState?: AgentState;

Defined in: packages/nexus-agents/src/agents/state-machine-types.ts:59

Initial state (defaults to ‘idle’)

maxErrorCount?
optional maxErrorCount?: number;

Defined in: packages/nexus-agents/src/agents/state-machine-types.ts:61

Maximum error count before permanent error state

maxHistorySize?
optional maxHistorySize?: number;

Defined in: packages/nexus-agents/src/agents/state-machine-types.ts:65

Maximum history entries to keep

trackHistory?
optional trackHistory?: boolean;

Defined in: packages/nexus-agents/src/agents/state-machine-types.ts:63

Enable transition history tracking


StateTransition

Defined in: packages/nexus-agents/src/agents/state-machine-types.ts:27

State transition metadata.

Properties

context?
optional context?: Record<string, unknown>;

Defined in: packages/nexus-agents/src/agents/state-machine-types.ts:37

Optional context data

event
event: StateTransitionEvent;

Defined in: packages/nexus-agents/src/agents/state-machine-types.ts:33

Event that triggered the transition

from
from: AgentState;

Defined in: packages/nexus-agents/src/agents/state-machine-types.ts:29

Previous state

timestamp
timestamp: string;

Defined in: packages/nexus-agents/src/agents/state-machine-types.ts:35

Timestamp of the transition

to
to: AgentState;

Defined in: packages/nexus-agents/src/agents/state-machine-types.ts:31

New state


StatusUpdateMessage

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:175

Status update message.

Properties

estimatedTimeRemaining?
optional estimatedTimeRemaining?: number;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:180

expertId
expertId: string;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:177

progress?
optional progress?: number;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:179

status
status: "failed" | "working" | "pending" | "submitted" | "reviewing" | "voted";

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:178

type
type: "status_update";

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:176


SubTask

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:26

A subtask broken down from the main task.

Properties

assignedRole?
optional assignedRole?: AgentRole;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:42

Assigned expert role (if any)

complexity
complexity: number;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:44

Estimated complexity (1-10)

dependencies
dependencies: string[];

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:36

Dependencies on other subtasks (by ID)

description
description: string;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:32

Description of what needs to be done

expectedOutput
expectedOutput: string;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:34

Expected output format or type

id
id: string;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:28

Unique subtask identifier

parentTaskId
parentTaskId: string;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:30

Parent task ID

priority
priority: SubtaskPriority;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:38

Priority level

requiredCapabilities
requiredCapabilities: string[];

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:46

Required capabilities for this subtask

status
status: SubtaskStatus;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:40

Current status


SynthesizedResult

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:126

Synthesis of multiple task results.

Properties

collaborationMetadata?
optional collaborationMetadata?: CollaborationMetadata;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:140

Collaboration metadata if collaborative synthesis was used (Issue #488)

combinedOutput
combinedOutput: string;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:128

Combined output from all results

conflicts
conflicts: Conflict[];

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:134

Any conflicts detected between results

qualityScore
qualityScore: number;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:136

Overall quality assessment

recommendations
recommendations: string[];

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:138

Recommendations for follow-up

resultSummaries
resultSummaries: ResultSummary[];

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:132

Individual result summaries

summary
summary: string;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:130

Summary of the synthesis process


SystemComponent

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:163

System component in architecture.

Properties

dependencies
dependencies: string[];

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:171

Dependencies

name
name: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:165

Component name

responsibilities
responsibilities: string[];

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:169

Responsibilities

type
type: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:167

Component type


TaskAnalysis

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:52

Result of task analysis.

Properties

approach
approach: string;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:66

Recommended approach

commitment?
optional commitment?: TaskCommitment;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:75

Commit-before-generate block (#1827). Forces the orchestrator to commit to a direction before dispatching workers, countering LLM mode-collapse toward safe defaults. Optional for backward compatibility with older analysis outputs.

complexity
complexity: number;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:56

Overall complexity score (1-10)

estimatedEffort
estimatedEffort: number;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:68

Estimated total effort in relative units

needsDecomposition
needsDecomposition: boolean;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:64

Whether task needs decomposition

requirements
requirements: string[];

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:60

Key requirements extracted from the task

risks
risks: string[];

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:62

Identified risks or challenges

taskId
taskId: string;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:54

Task ID being analyzed

taskType
taskType: string;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:58

Type of task (code, architecture, documentation, etc.)


TaskAssignmentMessage

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:105

Task assignment message sent to an expert.

Properties

deadline?
optional deadline?: string;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:111

expertId
expertId: string;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:107

previousResults?
optional previousResults?: TaskResult[];

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:110

sequencePosition?
optional sequencePosition?: number;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:109

task
task: Task;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:108

type
type: "task_assignment";

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:106


TaskCommitment

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:82

Orchestrator’s directional commitment, emitted before decomposition (#1827). Modeled after the frontend-design plugin’s Design Thinking pre-phase.

Properties

approach
approach: string;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:86

The non-obvious choice being made.

constraints
constraints: string[];

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:90

Hard limits: deadlines, scope boundaries, invariants.

differentiation
differentiation: string;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:88

What would make the output worse if solved by default patterns.

purpose
purpose: string;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:84

What this task is fundamentally about (one sentence).


TestingAnalysisResult

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:177

Testing analysis result from TestingExpert.

Extends

Properties

confidence
confidence: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:42

Confidence score (0-1)

Inherited from

ExpertOutput.confidence

content
content: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:34

Primary result content

Inherited from

ExpertOutput.content

coverage?
optional coverage?: CoverageMetrics;

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:183

Coverage metrics

modelUsed?
optional modelUsed?: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:44

Model used for this expert’s execution (Issue #817)

Inherited from

ExpertOutput.modelUsed

operationType
operationType: "generation" | "coverage_analysis" | "quality_assessment";

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:179

Operation type

quality?
optional quality?: TestQuality;

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:185

Test quality assessment

recommendations?
optional recommendations?: string[];

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:38

Recommendations or suggestions

Inherited from

ExpertOutput.recommendations

structuredData?
optional structuredData?: Record<string, unknown>;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:36

Structured data if applicable

Inherited from

ExpertOutput.structuredData

tests?
optional tests?: GeneratedTest[];

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:181

Generated tests

warnings?
optional warnings?: string[];

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:40

Warnings or issues found

Inherited from

ExpertOutput.warnings


TestingExpertOptions

Defined in: packages/nexus-agents/src/agents/experts/testing-expert.ts:44

Configuration options for TestingExpert.

Extends

Properties

additionalCapabilities?
optional additionalCapabilities?: AgentCapability[];

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:26

Custom capability extensions

Inherited from

ExpertOptions.additionalCapabilities

enableHeuristics?
optional enableHeuristics?: boolean;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:24

Enable domain-specific heuristics

Inherited from

ExpertOptions.enableHeuristics

framework?
optional framework?: "vitest" | "jest" | "playwright" | "mocha" | "cypress";

Defined in: packages/nexus-agents/src/agents/experts/testing-expert.ts:46

Preferred testing framework

generateFactories?
optional generateFactories?: boolean;

Defined in: packages/nexus-agents/src/agents/experts/testing-expert.ts:54

Generate test data factories

includeMocking?
optional includeMocking?: boolean;

Defined in: packages/nexus-agents/src/agents/experts/testing-expert.ts:50

Include mocking strategies

maxTokens?
optional maxTokens?: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:22

Maximum tokens for responses

Inherited from

ExpertOptions.maxTokens

systemPromptOverride?
optional systemPromptOverride?: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:18

Custom system prompt override

Inherited from

ExpertOptions.systemPromptOverride

targetCoverage?
optional targetCoverage?: number;

Defined in: packages/nexus-agents/src/agents/experts/testing-expert.ts:48

Target coverage percentage

temperature?
optional temperature?: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-base-types.ts:20

Temperature for completions (domain-specific default if not set)

Inherited from

ExpertOptions.temperature

testStyle?
optional testStyle?: "behavioral" | "tdd" | "bdd";

Defined in: packages/nexus-agents/src/agents/experts/testing-expert.ts:52

Test style preference


TestQuality

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:223

Test quality assessment.

Properties

assertionQuality
assertionQuality: "good" | "fair" | "poor";

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:229

Assertion quality

isolation
isolation: "good" | "fair" | "poor";

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:227

Test isolation

issues
issues: string[];

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:231

Issues found

score
score: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:225

Overall score (0-100)


ThinkerOutput

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:128

Thinker’s analysis output.

Properties

approach
readonly approach: string;

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:132

Execution approach/plan

considerations
readonly considerations: string[];

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:134

Considerations and edge cases

problemAnalysis
readonly problemAnalysis: string;

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:130

Problem analysis

successCriteria
readonly successCriteria: string[];

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:136

Success criteria


TreeStatistics

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:131

Statistics for a reasoning tree.

Properties

activeNodes
readonly activeNodes: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:135

Number of currently active nodes

avgBranchingFactor
readonly avgBranchingFactor: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:147

Average branching factor

avgConfidence
readonly avgConfidence: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:141

Average node confidence

avgQualityScore
readonly avgQualityScore: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:139

Average node quality score

conclusionCount
readonly conclusionCount: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:143

Number of conclusion nodes

maxDepth
readonly maxDepth: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:137

Maximum depth reached

totalNodes
readonly totalNodes: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:133

Total number of nodes in the tree

totalTokensUsed
readonly totalTokensUsed: number;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:145

Total tokens used across all nodes


TrinityConfig

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:170

Configuration for TRINITY coordinator.

Properties

includeHistory?
readonly optional includeHistory?: boolean;

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:176

Whether to include detailed phase history

maxIterations?
readonly optional maxIterations?: number;

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:172

Maximum verification iterations before giving up

roleConfigs?
readonly optional roleConfigs?: Partial<Record<TrinityRole, Partial<TrinityRoleConfig>>>;

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:178

Custom role configurations

timeoutMs?
readonly optional timeoutMs?: number;

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:174

Timeout for entire coordination in ms


TrinityExecuteOptions

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:253

Options for executing TRINITY coordination.

Properties

agent
readonly agent: IAgent;

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:255

task
readonly task: Task;

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:254


TrinityPhaseResult

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:105

Result from a single TRINITY phase.

Properties

durationMs
readonly durationMs: number;

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:113

Duration in milliseconds

output
readonly output: string;

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:111

Output from the phase

phase
readonly phase: TrinityPhase;

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:107

Which phase produced this result

role
readonly role: TrinityRole;

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:109

Role that executed this phase

tokensMeasured?
readonly optional tokensMeasured?: boolean;

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:124

Whether tokensUsed is a measurement (#4743).

false means the adapter reported no usage, so tokensUsed is a placeholder zero rather than a count. Absent means the producer predates the distinction — unknown, not measured. Additive and optional, so no existing reader breaks.

tokensUsed
readonly tokensUsed: number;

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:115

Tokens used. Meaningful only when tokensMeasured is not false.


TrinityResult

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:194

Result of TRINITY coordination.

Properties

finalOutput
readonly finalOutput: string;

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:198

Final output after all phases

history
readonly history: TrinityPhaseResult[];

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:210

Phase execution history

iterations
readonly iterations: number;

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:206

Number of think-work-verify iterations

stopReason
readonly stopReason: "error" | "timeout" | "verified" | "max_iterations";

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:212

Stop reason

success
readonly success: boolean;

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:196

Whether coordination succeeded

thinkerOutput
readonly thinkerOutput: ThinkerOutput;

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:200

Thinker’s analysis

totalDurationMs
readonly totalDurationMs: number;

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:208

Total duration in milliseconds

verifierOutput
readonly verifierOutput: VerifierOutput;

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:204

Verifier’s final assessment

workerOutput
readonly workerOutput: WorkerOutput;

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:202

Worker’s implementation


TrinityRoleConfig

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:23

Configuration for a TRINITY role.

Properties

maxTokens
readonly maxTokens: number;

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:31

Maximum tokens for response

role
readonly role: TrinityRole;

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:25

Role identifier

systemPrompt
readonly systemPrompt: string;

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:27

System prompt for this role

temperature
readonly temperature: number;

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:29

Temperature for completions


VerifierOutput

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:152

Verifier’s evaluation output.

Properties

correctnessCheck
readonly correctnessCheck: string;

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:156

Correctness assessment

issuesFound
readonly issuesFound: string[];

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:160

Issues found

qualityCheck
readonly qualityCheck: string;

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:158

Quality assessment

recommendations
readonly recommendations: string[];

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:162

Recommendations

verdict
readonly verdict: "pass" | "fail";

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:154

Pass or fail verdict


VoteMessage

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:164

Vote message for consensus protocol.

Properties

conditions?
optional conditions?: string[];

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:169

decision
decision: VoteDecision;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:167

expertId
expertId: string;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:166

reasoning
reasoning: string;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:168

type
type: "vote";

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:165


Vulnerability

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:85

Represents a security vulnerability.

Properties

cweId?
optional cweId?: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:99

CWE reference if applicable

description
description: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:93

Description of the vulnerability

id
id: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:87

Unique vulnerability ID

location?
optional location?: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:95

Affected location

remediation
remediation: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:97

Remediation steps

severity
severity: "critical" | "info" | "high" | "low" | "medium";

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:89

Severity level

type
type: string;

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:91

Vulnerability type (OWASP category)


WaveExecutionResult

Defined in: packages/nexus-agents/src/agents/wave-scheduler-types.ts:127

Final result of the full wave execution.

Properties

aborted
readonly aborted: boolean;

Defined in: packages/nexus-agents/src/agents/wave-scheduler-types.ts:137

Whether execution was aborted early (budget exceeded or failure).

abortReason?
readonly optional abortReason?: string;

Defined in: packages/nexus-agents/src/agents/wave-scheduler-types.ts:139

Reason for abort, if aborted.

allResults
readonly allResults: readonly WaveTaskResult[];

Defined in: packages/nexus-agents/src/agents/wave-scheduler-types.ts:131

All task results flat.

totalDurationMs
readonly totalDurationMs: number;

Defined in: packages/nexus-agents/src/agents/wave-scheduler-types.ts:135

Total duration in ms.

totalTokensUsed
readonly totalTokensUsed: number;

Defined in: packages/nexus-agents/src/agents/wave-scheduler-types.ts:133

Total estimated tokens consumed across all waves.

waves
readonly waves: readonly WaveResult[];

Defined in: packages/nexus-agents/src/agents/wave-scheduler-types.ts:129

Results organized by wave.


WaveResult

Defined in: packages/nexus-agents/src/agents/wave-scheduler-types.ts:113

Result of executing a single wave.

Properties

durationMs
readonly durationMs: number;

Defined in: packages/nexus-agents/src/agents/wave-scheduler-types.ts:121

Total duration of this wave in ms.

results
readonly results: readonly WaveTaskResult[];

Defined in: packages/nexus-agents/src/agents/wave-scheduler-types.ts:117

Results of all tasks in this wave.

totalTokens
readonly totalTokens: number;

Defined in: packages/nexus-agents/src/agents/wave-scheduler-types.ts:119

Total estimated tokens consumed by this wave.

waveIndex
readonly waveIndex: number;

Defined in: packages/nexus-agents/src/agents/wave-scheduler-types.ts:115

The wave index (0-based).


WaveSchedulerConfig

Defined in: packages/nexus-agents/src/agents/wave-scheduler-types.ts:19

Configuration for the wave scheduler.

Properties

abortOnFailure
readonly abortOnFailure: boolean;

Defined in: packages/nexus-agents/src/agents/wave-scheduler-types.ts:35

Whether to abort remaining waves on first task failure. Default: false.

maxConcurrency
readonly maxConcurrency: number;

Defined in: packages/nexus-agents/src/agents/wave-scheduler-types.ts:21

Maximum number of tasks to execute concurrently in one wave. Default: 4.

maxOutputChars
readonly maxOutputChars: number;

Defined in: packages/nexus-agents/src/agents/wave-scheduler-types.ts:23

Maximum output length (chars) per task result. Default: 2000.

maxTotalTokens
readonly maxTotalTokens: number;

Defined in: packages/nexus-agents/src/agents/wave-scheduler-types.ts:33

Maximum total token budget across all waves. 0 = unlimited. Default: 0.

Enforced against the SUM OF ESTIMATES described on WaveTaskResult.estimatedTokens, which excludes input tokens and counts a failed task as zero. Both errors run in the same direction: the budget believes it has more headroom than it does, so enabling this does not give you a reliable spend cap (#4761).

onWaveComplete?
readonly optional onWaveComplete?: (waveIndex, results, cumulativeTokens) => Promise<void>;

Defined in: packages/nexus-agents/src/agents/wave-scheduler-types.ts:39

Optional callback invoked after each wave completes. Used for checkpointing.

Parameters
waveIndex

number

results

readonly WaveTaskResult[]

cumulativeTokens

number

Returns

Promise<void>

taskTimeoutMs
readonly taskTimeoutMs: number;

Defined in: packages/nexus-agents/src/agents/wave-scheduler-types.ts:37

Timeout per individual task in ms. Default: 60000.


WaveTask

Defined in: packages/nexus-agents/src/agents/wave-scheduler-types.ts:65

A task to be executed in a wave.

Type Parameters

T

T = unknown

Properties

dependencies
readonly dependencies: readonly string[];

Defined in: packages/nexus-agents/src/agents/wave-scheduler-types.ts:73

IDs of tasks that must complete before this one can start.

description
readonly description: string;

Defined in: packages/nexus-agents/src/agents/wave-scheduler-types.ts:69

Human-readable description of what this task does.

id
readonly id: string;

Defined in: packages/nexus-agents/src/agents/wave-scheduler-types.ts:67

Unique identifier for this task.

input
readonly input: T;

Defined in: packages/nexus-agents/src/agents/wave-scheduler-types.ts:71

The input data for this task.


WaveTaskResult

Defined in: packages/nexus-agents/src/agents/wave-scheduler-types.ts:79

Result of a single task execution.

Properties

durationMs
readonly durationMs: number;

Defined in: packages/nexus-agents/src/agents/wave-scheduler-types.ts:105

Duration of this task in ms.

error?
readonly optional error?: string;

Defined in: packages/nexus-agents/src/agents/wave-scheduler-types.ts:107

Error message if task failed.

estimatedTokens
readonly estimatedTokens: number;

Defined in: packages/nexus-agents/src/agents/wave-scheduler-types.ts:103

Rough token estimate for this task, derived as outputChars / 4 (#4761).

NOT a measurement, and specifically NOT comparable to ResultMetadata.tokensUsed:

  • Input is not counted. Prompt and context usually dominate an agent task’s spend, so a large prompt with a terse answer looks nearly free.
  • A failed task reports 0, however long it ran before throwing.

Real usage is not available here: WaveTaskExecutor returns a bare string, so the scheduler has nothing better to sum. Treat this as a coarse output-size signal, not a spend figure.

originalLength
readonly originalLength: number;

Defined in: packages/nexus-agents/src/agents/wave-scheduler-types.ts:89

Original output length before truncation.

output
readonly output: string;

Defined in: packages/nexus-agents/src/agents/wave-scheduler-types.ts:85

The output text (truncated to maxOutputChars).

success
readonly success: boolean;

Defined in: packages/nexus-agents/src/agents/wave-scheduler-types.ts:83

Whether the task completed successfully.

taskId
readonly taskId: string;

Defined in: packages/nexus-agents/src/agents/wave-scheduler-types.ts:81

Task ID this result belongs to.

truncated
readonly truncated: boolean;

Defined in: packages/nexus-agents/src/agents/wave-scheduler-types.ts:87

Whether the output was truncated.


WorkChunk

Defined in: packages/nexus-agents/src/agents/wave-scheduler-types.ts:149

A chunk of work produced by auto-chunking.

Properties

id
readonly id: string;

Defined in: packages/nexus-agents/src/agents/wave-scheduler-types.ts:151

Unique ID for this chunk.

items
readonly items: readonly string[];

Defined in: packages/nexus-agents/src/agents/wave-scheduler-types.ts:155

Items in this chunk (e.g., file paths).

scope
readonly scope: string;

Defined in: packages/nexus-agents/src/agents/wave-scheduler-types.ts:153

Scope description (e.g., directory path).


WorkerOutput

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:140

Worker’s implementation output.

Properties

deviations
readonly deviations: string[];

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:146

Deviations from plan

implementation
readonly implementation: string;

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:142

The actual implementation/content

questions
readonly questions: string[];

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:148

Questions or blockers

stepsCompleted
readonly stepsCompleted: string[];

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:144

Steps completed

Type Aliases

ActivationStrategy

type ActivationStrategy = "ucb" | "greedy" | "diverse" | "adaptive";

Defined in: packages/nexus-agents/src/agents/reasoning/forest-config-types.ts:25

Strategy for selecting which nodes to activate.

  • ucb: Upper Confidence Bound (exploration/exploitation balance)
  • greedy: Always activate highest-scoring nodes
  • diverse: Prioritize diversity across trees
  • adaptive: Dynamically adjust based on progress

AgentStopReason

type AgentStopReason = "agent-stopped" | "turn-budget" | "tool-error" | "cancelled";

Defined in: packages/nexus-agents/src/agents/agentic/types.ts:72

Why the agent loop stopped.

  • agent-stopped: model emitted no further tool calls — natural end
  • turn-budget: hit turnBudget before the model finished
  • tool-error: onToolCall threw; harness’s responsibility to grade
  • cancelled: external AbortSignal fired

AggregationStrategy

type AggregationStrategy = "merge" | "select_best" | "consensus" | "sequential_chain";

Defined in: packages/nexus-agents/src/agents/collaboration/aggregator-types.ts:19

Aggregation strategy types.


ArchitectureStyle

type ArchitectureStyle = 
  | "layered"
  | "microservices"
  | "event_driven"
  | "hexagonal"
  | "clean"
  | "cqrs"
  | "ddd";

Defined in: packages/nexus-agents/src/agents/experts/architecture-expert.ts:52

Architecture style options.


BuiltInExpertType

type BuiltInExpertType = 
  | "code"
  | "architecture"
  | "security"
  | "documentation"
  | "testing"
  | "devops"
  | "research"
  | "pm"
  | "ux"
  | "infrastructure"
  | "qa"
  | "data-visualization";

Defined in: packages/nexus-agents/src/agents/experts/expert-config.ts:70

Built-in expert type identifiers.


CollaborationMessage

type CollaborationMessage = 
  | TaskAssignmentMessage
  | ResultSubmissionMessage
  | ReviewRequestMessage
  | ReviewResponseMessage
  | FeedbackMessage
  | VoteMessage
  | StatusUpdateMessage;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:93

Collaboration message types for inter-expert communication.


CollaborationPattern

type CollaborationPattern = 
  | "sequential"
  | "parallel"
  | "review"
  | "consensus"
  | "reflexion"
  | "aegean"
  | "self-refine"
  | "self-debug";

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:36

Collaboration pattern types.

  • sequential: Experts work in order, passing results forward
  • parallel: Experts work simultaneously on the same task
  • review: One expert reviews another’s work
  • consensus: Voting-based decision making
  • reflexion: Multi-agent reflexion with persona-based critics (arxiv:2512.20845)
  • aegean: Byzantine-fault-tolerant consensus (arxiv:2512.20184)
  • self-refine: Iterative refinement with self-feedback (arxiv:2303.17651)
  • self-debug: Automatic error detection and repair (arxiv:2304.05128)

ConflictResolver

type ConflictResolver = (conflict, result1, result2) => "expert1" | "expert2" | "merged";

Defined in: packages/nexus-agents/src/agents/collaboration/aggregator-types.ts:34

Conflict resolver function type.

Parameters

conflict

ResultConflict

result1

ExpertResult

result2

ExpertResult

Returns

"expert1" | "expert2" | "merged"


ContentPriority

type ContentPriority = typeof ContentPriority[keyof typeof ContentPriority];

Defined in: packages/nexus-agents/src/agents/context-manager-types.ts:13

Priority levels for context content. Higher priority content is retained longer during pruning.


CrossTreeStrategy

type CrossTreeStrategy = "none" | "conclusions" | "insights" | "full";

Defined in: packages/nexus-agents/src/agents/reasoning/forest-config-types.ts:40

Strategy for sharing information across trees.

  • none: No cross-tree sharing
  • conclusions: Share only conclusions
  • insights: Share conclusions and intermediate insights
  • full: Share all relevant information

ExpertCollaborationPatternType

type ExpertCollaborationPatternType = typeof ExpertCollaborationPattern[keyof typeof ExpertCollaborationPattern];

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:28

Agents exports - Agent framework, Orchestrator, Experts Split from index.ts for file size compliance (Issue #285)


ExpertDomain

type ExpertDomain = "code" | "security" | "architecture" | "testing" | "documentation";

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:28

Expert domain categories.


ExplorationEventType

type ExplorationEventType = 
  | "tree_created"
  | "node_created"
  | "node_activated"
  | "node_deactivated"
  | "node_completed"
  | "node_pruned"
  | "path_scored"
  | "cross_tree_share"
  | "conclusion_reached"
  | "tree_completed"
  | "forest_converging"
  | "forest_completed";

Defined in: packages/nexus-agents/src/agents/reasoning/forest-result-types.ts:89

Types of exploration events.


FailureClassification

type FailureClassification = {
  archetype?: FailureArchetype;
  confidence?: number;
  kind: "transient" | "permanent";
  source: "transport" | "archetype" | "default";
};

Defined in: packages/nexus-agents/src/agents/experts/expert-recovery.ts:115

The outcome of classifying a single failed execution attempt.

  • transient: retry (transport error, or a recoverable behavioral archetype)
  • permanent: fail closed (cancelled, non-retryable, or a terminal archetype)

Properties

archetype?
optional archetype?: FailureArchetype;

Defined in: packages/nexus-agents/src/agents/experts/expert-recovery.ts:118

confidence?
optional confidence?: number;

Defined in: packages/nexus-agents/src/agents/experts/expert-recovery.ts:119

kind
kind: "transient" | "permanent";

Defined in: packages/nexus-agents/src/agents/experts/expert-recovery.ts:116

source
source: "transport" | "archetype" | "default";

Defined in: packages/nexus-agents/src/agents/experts/expert-recovery.ts:117


ForestId

type ForestId = string;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-node-types.ts:30

Unique identifier for a forest (collection of trees).


ForestPruningStrategy

type ForestPruningStrategy = "none" | "score" | "depth" | "combined";

Defined in: packages/nexus-agents/src/agents/reasoning/forest-config-types.ts:55

Strategy for pruning low-quality branches.

  • none: No pruning
  • score: Prune nodes below score threshold
  • depth: Prune based on depth limits
  • combined: Use both score and depth criteria

ForestState

type ForestState = "initializing" | "exploring" | "converging" | "completed" | "timeout";

Defined in: packages/nexus-agents/src/agents/reasoning/forest-state-types.ts:28

State of a forest of reasoning trees.

  • initializing: Forest is being set up
  • exploring: Actively exploring trees
  • converging: Trees are converging on solution(s)
  • completed: Forest has finished exploration
  • timeout: Exploration ended due to timeout

NodeId

type NodeId = string;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-node-types.ts:20

Unique identifier for a reasoning node.


NodeState

type NodeState = "pending" | "active" | "completed" | "pruned" | "error";

Defined in: packages/nexus-agents/src/agents/reasoning/forest-node-types.ts:45

State of a reasoning node in its lifecycle.

  • pending: Node created but not yet processed
  • active: Node currently being explored/evaluated
  • completed: Node exploration finished successfully
  • pruned: Node was pruned due to low score or depth limit
  • error: Node exploration failed with an error

PruningStrategy

type PruningStrategy = typeof PruningStrategy[keyof typeof PruningStrategy];

Defined in: packages/nexus-agents/src/agents/context-pruner.ts:41

Strategy for pruning context when budget is exceeded.


QualityAttribute

type QualityAttribute = 
  | "performance"
  | "scalability"
  | "maintainability"
  | "security"
  | "reliability"
  | "testability";

Defined in: packages/nexus-agents/src/agents/experts/architecture-expert.ts:58

Quality attributes for architecture decisions.


QualityScorer

type QualityScorer = (results, aggregatedOutput) => number;

Defined in: packages/nexus-agents/src/agents/collaboration/aggregator-types.ts:43

Quality scorer function type.

Parameters

results

ExpertResult[]

aggregatedOutput

unknown

Returns

number


ReasoningStepType

type ReasoningStepType = 
  | "hypothesis"
  | "inference"
  | "decomposition"
  | "synthesis"
  | "verification"
  | "conclusion";

Defined in: packages/nexus-agents/src/agents/reasoning/forest-node-types.ts:62

Type of reasoning step represented by a node.

  • hypothesis: Initial hypothesis or assumption
  • inference: Logical deduction from parent node(s)
  • decomposition: Breaking down a complex problem
  • synthesis: Combining multiple reasoning paths
  • verification: Validating a previous step
  • conclusion: Final answer or decision

SecurityFocusArea

type SecurityFocusArea = 
  | "authentication"
  | "authorization"
  | "input_validation"
  | "cryptography"
  | "injection"
  | "secrets"
  | "dependencies";

Defined in: packages/nexus-agents/src/agents/experts/security-expert.ts:54

Security focus areas for targeted analysis.


SessionEvent

type SessionEvent = 
  | {
  status: SessionStatus;
  type: "status_change";
}
  | {
  expertId: string;
  type: "expert_joined";
}
  | {
  expertId: string;
  result: TaskResult;
  type: "result_submitted";
}
  | {
  approved: boolean;
  reviewerId: string;
  type: "review_completed";
}
  | {
  decision: string;
  expertId: string;
  type: "vote_received";
}
  | {
  expertId?: string;
  type: "timeout";
}
  | {
  error: Error;
  type: "error";
};

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-session-helpers.ts:29

Session event types for callbacks.


SessionStatus

type SessionStatus = 
  | "pending"
  | "in_progress"
  | "awaiting_review"
  | "voting"
  | "finalizing"
  | "completed"
  | "failed"
  | "timed_out";

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:49

Session status during collaboration lifecycle.


StateChangeCallback

type StateChangeCallback = (transition) => void;

Defined in: packages/nexus-agents/src/agents/state-machine-types.ts:43

Callback for state change events.

Parameters

transition

StateTransition

Returns

void


StateTransitionEvent

type StateTransitionEvent = 
  | "task_assigned"
  | "plan_completed"
  | "needs_input"
  | "task_completed"
  | "failure"
  | "input_received"
  | "recovered";

Defined in: packages/nexus-agents/src/agents/state-machine-types.ts:15

State transition event types.


SubtaskPriority

type SubtaskPriority = "critical" | "high" | "medium" | "low";

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:16

Subtask priority levels.


SubtaskStatus

type SubtaskStatus = "pending" | "assigned" | "in_progress" | "completed" | "failed";

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:21

Subtask status.


TerminationReason

type TerminationReason = 
  | "solution_found"
  | "convergence"
  | "max_time"
  | "max_tokens"
  | "max_depth"
  | "no_progress"
  | "error";

Defined in: packages/nexus-agents/src/agents/reasoning/forest-result-types.ts:26

Termination reason for forest exploration.


TransitionErrorCallback

type TransitionErrorCallback = (currentState, attemptedEvent, error) => void;

Defined in: packages/nexus-agents/src/agents/state-machine-types.ts:48

Error callback for invalid transitions.

Parameters

currentState

AgentState

attemptedEvent

StateTransitionEvent

error

AgentError

Returns

void


TreeId

type TreeId = string;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-node-types.ts:25

Unique identifier for a reasoning tree.


TreeState

type TreeState = "growing" | "paused" | "completed" | "abandoned";

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:26

State of a reasoning tree.

  • growing: Tree is actively being explored
  • paused: Tree exploration temporarily paused
  • completed: Tree has reached conclusion(s)
  • abandoned: Tree was abandoned (low quality or pruned)

TrinityPhase

type TrinityPhase = "thinking" | "working" | "verifying" | "complete";

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:102

Phase of TRINITY coordination.


TrinityRole

type TrinityRole = "thinker" | "worker" | "verifier";

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:20

TRINITY-specific roles.


VoteDecision

type VoteDecision = "approve" | "reject" | "abstain";

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-types.ts:62

Vote decision options for consensus protocol.


WaveTaskExecutor

type WaveTaskExecutor<T> = (task) => Promise<string>;

Defined in: packages/nexus-agents/src/agents/wave-scheduler-types.ts:162

Executor function that processes a single WaveTask. Returns the output string (which will be truncated by the scheduler).

Type Parameters

T

T = unknown

Parameters

task

WaveTask<T>

Returns

Promise<string>

Variables

ActivationStrategySchema

const ActivationStrategySchema: ZodEnum<{
  adaptive: "adaptive";
  diverse: "diverse";
  greedy: "greedy";
  ucb: "ucb";
}>;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-config-types.ts:30

Schema for ActivationStrategy validation.


AgentMessageSchema

const AgentMessageSchema: ZodObject<{
  from: ZodString;
  id: ZodString;
  payload: ZodUnknown;
  timestamp: ZodString;
  to: ZodString;
  type: ZodEnum<{
     feedback: "feedback";
     query: "query";
     result: "result";
     status: "status";
     task: "task";
  }>;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/agent-schemas.ts:54

Zod schema for validating AgentMessage objects.


BaseAgentOptionsSchema

const BaseAgentOptionsSchema: ZodObject<{
  capabilities: ZodArray<ZodEnum<{
     code_generation: "code_generation";
     code_review: "code_review";
     collaboration: "collaboration";
     delegation: "delegation";
     research: "research";
     task_execution: "task_execution";
     tool_use: "tool_use";
  }>>;
  contextPruning: ZodOptional<ZodObject<{
     enabled: ZodOptional<ZodBoolean>;
     maxTokens: ZodOptional<ZodNumber>;
     reserveTokens: ZodOptional<ZodNumber>;
     strategy: ZodOptional<ZodEnum<{
        hierarchical: "hierarchical";
        lowest_priority: "lowest_priority";
        oldest_first: "oldest_first";
        priority_weighted_age: "priority_weighted_age";
        semantic: "semantic";
        sliding_window: "sliding_window";
        summarize: "summarize";
     }>>;
     triggerThreshold: ZodOptional<ZodNumber>;
  }, $strip>>;
  id: ZodString;
  maxTokens: ZodOptional<ZodNumber>;
  role: ZodEnum<{
     architecture_expert: "architecture_expert";
     code_expert: "code_expert";
     custom: "custom";
     data_visualization_expert: "data_visualization_expert";
     devops_expert: "devops_expert";
     documentation_expert: "documentation_expert";
     infrastructure_expert: "infrastructure_expert";
     orchestrator: "orchestrator";
     pm_expert: "pm_expert";
     qa_expert: "qa_expert";
     research_expert: "research_expert";
     security_expert: "security_expert";
     testing_expert: "testing_expert";
     thinker: "thinker";
     ux_expert: "ux_expert";
     verifier: "verifier";
     worker: "worker";
  }>;
  systemPrompt: ZodOptional<ZodString>;
  temperature: ZodOptional<ZodNumber>;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/agent-schemas.ts:93

Zod schema for validating BaseAgentOptions.


BestSolutionSchema

const BestSolutionSchema: ZodObject<{
  combinedScore: ZodNumber;
  conclusionNode: ZodObject<{
     activationScore: ZodNumber;
     children: ZodArray<ZodString>;
     confidence: ZodNumber;
     content: ZodString;
     createdAt: ZodNumber;
     depth: ZodNumber;
     estimatedValue: ZodNumber;
     id: ZodString;
     isActive: ZodBoolean;
     metadata: ZodObject<{
        crossReferences: ZodOptional<ZodArray<ZodString>>;
        custom: ZodOptional<ZodRecord<ZodString, ZodUnknown>>;
        generationTimeMs: ZodOptional<ZodNumber>;
        source: ZodOptional<ZodString>;
        tokensUsed: ZodOptional<ZodNumber>;
     }, $strip>;
     parentId: ZodNullable<ZodString>;
     qualityScore: ZodNumber;
     state: ZodEnum<{
        active: "active";
        completed: "completed";
        error: "error";
        pending: "pending";
        pruned: "pruned";
     }>;
     stepType: ZodEnum<{
        conclusion: "conclusion";
        decomposition: "decomposition";
        hypothesis: "hypothesis";
        inference: "inference";
        synthesis: "synthesis";
        verification: "verification";
     }>;
     treeId: ZodString;
     updatedAt: ZodNumber;
  }, $strip>;
  confidence: ZodNumber;
  path: ZodArray<ZodString>;
  qualityScore: ZodNumber;
  treeId: ZodString;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-result-types.ts:73

Schema for BestSolution validation.


BUILT_IN_EXPERTS

const BUILT_IN_EXPERTS: Readonly<Record<BuiltInExpertType, ExpertConfig>>;

Defined in: packages/nexus-agents/src/agents/experts/expert-config.ts:195

Built-in expert configurations. These provide sensible defaults for common expert types.

MIRROR of agents/<name>-expert.md, which is repo-only and NOT shipped — package.json#files lists dist, two src/ asset dirs and the README, so an installed package has no agents/ directory. The runtime reads this constant, never those files, which is why the omission is correct rather than a bug (#5143).

GATED BY scripts/generate-agents-index.ts --check, which fails when an expert exists in one place and not the other. Without that gate this constant and its source documents would drift silently, and the drift would only be visible to someone reading both.


BuiltInExpertTypeSchema

const BuiltInExpertTypeSchema: ZodEnum<{
  architecture: "architecture";
  code: "code";
  data-visualization: "data-visualization";
  devops: "devops";
  documentation: "documentation";
  infrastructure: "infrastructure";
  pm: "pm";
  qa: "qa";
  research: "research";
  security: "security";
  testing: "testing";
  ux: "ux";
}>;

Defined in: packages/nexus-agents/src/agents/experts/expert-config.ts:165

Zod schema for BuiltInExpertType.

MUST stay in lockstep with the BuiltInExpertType type union above. Tested by BuiltInExpertTypeSchema accepts every literal in BuiltInExpertType in expert-config.test.ts to prevent drift (#2338).


CodeChangeSchema

const CodeChangeSchema: ZodObject<{
  description: ZodString;
  file: ZodString;
  lineRange: ZodOptional<ZodObject<{
     end: ZodNumber;
     start: ZodNumber;
  }, $strip>>;
  modified: ZodString;
  original: ZodOptional<ZodString>;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:292

Code change schema.


CollaborationConfigSchema

const CollaborationConfigSchema: ZodObject<{
  experts: ZodArray<ZodString>;
  maxRetries: ZodOptional<ZodNumber>;
  minVotes: ZodOptional<ZodNumber>;
  pattern: ZodEnum<{
     aegean: "aegean";
     consensus: "consensus";
     parallel: "parallel";
     reflexion: "reflexion";
     review: "review";
     self-debug: "self-debug";
     self-refine: "self-refine";
     sequential: "sequential";
  }>;
  requireUnanimous: ZodOptional<ZodBoolean>;
  sessionId: ZodString;
  task: ZodObject<{
     constraints: ZodOptional<ZodObject<{
        allowedTools: ZodOptional<ZodArray<ZodString>>;
        maxDuration: ZodOptional<ZodNumber>;
        maxTokens: ZodOptional<ZodNumber>;
        outputFormat: ZodOptional<ZodEnum<{
           json: "json";
           markdown: "markdown";
           text: "text";
        }>>;
     }, $strip>>;
     context: ZodRecord<ZodString, ZodUnknown>;
     description: ZodString;
     id: ZodString;
     priority: ZodOptional<ZodNumber>;
  }, $strip>;
  timeout: ZodOptional<ZodNumber>;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-schemas.ts:45

Zod schema for CollaborationConfig.


CollaborationPatternSchema

const CollaborationPatternSchema: ZodEnum<{
  aegean: "aegean";
  consensus: "consensus";
  parallel: "parallel";
  reflexion: "reflexion";
  review: "review";
  self-debug: "self-debug";
  self-refine: "self-refine";
  sequential: "sequential";
}>;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-schemas.ts:12

Zod schema for CollaborationPattern.


ContentPriority

const ContentPriority: {
  ACTIVE: 60;
  EPHEMERAL: 20;
  HISTORY: 40;
  SYSTEM: 100;
  TASK: 80;
};

Defined in: packages/nexus-agents/src/agents/context-manager-types.ts:13

Priority levels for context content. Higher priority content is retained longer during pruning.

Type Declaration

ACTIVE
readonly ACTIVE: 60 = 60;

Active working content (recent code, research)

EPHEMERAL
readonly EPHEMERAL: 20 = 20;

Ephemeral content (debug logs, temp data)

HISTORY
readonly HISTORY: 40 = 40;

Historical context (older messages, results)

SYSTEM
readonly SYSTEM: 100 = 100;

System instructions - highest priority, never pruned

TASK
readonly TASK: 80 = 80;

Current task description and requirements


ContextBudgetSchema

const ContextBudgetSchema: ZodObject<{
  active: ZodNumber;
  reserved: ZodNumber;
  system: ZodNumber;
  task: ZodNumber;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/context-manager-types.ts:56

Zod schema for ContextBudget validation.


ContextManagerConfigSchema

const ContextManagerConfigSchema: ZodObject<{
  budget: ZodOptional<ZodObject<{
     active: ZodNumber;
     reserved: ZodNumber;
     system: ZodNumber;
     task: ZodNumber;
  }, $strip>>;
  maxTokens: ZodNumber;
  warningThreshold: ZodOptional<ZodNumber>;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/context-manager-types.ts:111

Schema for ContextManagerConfig validation.


ContextPrunerConfigSchema

const ContextPrunerConfigSchema: ZodObject<{
  autoTriggerThreshold: ZodOptional<ZodNumber>;
  defaultStrategy: ZodOptional<ZodEnum<{
     hierarchical: "hierarchical";
     lowest_priority: "lowest_priority";
     oldest_first: "oldest_first";
     priority_weighted_age: "priority_weighted_age";
     semantic: "semantic";
     sliding_window: "sliding_window";
     summarize: "summarize";
  }>>;
  minItemsPerCategory: ZodOptional<ZodNumber>;
  protectedPriority: ZodOptional<ZodNumber>;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/context-pruner.ts:64

Agents exports - Agent framework, Orchestrator, Experts Split from index.ts for file size compliance (Issue #285)


CoverageMetricsSchema

const CoverageMetricsSchema: ZodObject<{
  branch: ZodNumber;
  function: ZodNumber;
  line: ZodNumber;
  statement: ZodNumber;
  uncoveredAreas: ZodOptional<ZodArray<ZodString>>;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:319

Coverage metrics schema.


CrossTreeInfoSchema

const CrossTreeInfoSchema: ZodObject<{
  failurePatterns: ZodArray<ZodObject<{
     avgFailureScore: ZodNumber;
     occurrences: ZodNumber;
     pattern: ZodString;
  }, $strip>>;
  sharedConclusions: ZodArray<ZodObject<{
     confidence: ZodNumber;
     content: ZodString;
     qualityScore: ZodNumber;
     sourceNodeId: ZodString;
     sourceTreeId: ZodString;
  }, $strip>>;
  sharedInsights: ZodArray<ZodObject<{
     content: ZodString;
     relevance: ZodNumber;
     sourceNodeId: ZodString;
     sourceTreeId: ZodString;
  }, $strip>>;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-types.ts:119

Schema for CrossTreeInfo validation.


CrossTreeStrategySchema

const CrossTreeStrategySchema: ZodEnum<{
  conclusions: "conclusions";
  full: "full";
  insights: "insights";
  none: "none";
}>;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-config-types.ts:45

Schema for CrossTreeStrategy validation.


DEFAULT_ACTIVATION_OPTIONS

const DEFAULT_ACTIVATION_OPTIONS: ActivationOptions;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-config-types.ts:204

Default activation options.


DEFAULT_BUDGET

const DEFAULT_BUDGET: ContextBudget;

Defined in: packages/nexus-agents/src/agents/context-manager-types.ts:46

Default budget allocation percentages.


DEFAULT_FOREST_CONFIG

const DEFAULT_FOREST_CONFIG: ForestConfig;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-config-types.ts:154

Default Forest-of-Thought configuration.


DEFAULT_MAX_RETRIES

const DEFAULT_MAX_RETRIES: 2 = 2;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-schemas.ts:130

Default retry counts.


DEFAULT_PATH_SCORING_OPTIONS

const DEFAULT_PATH_SCORING_OPTIONS: PathScoringOptions;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:116

Default path scoring options.


DEFAULT_TIMEOUTS

const DEFAULT_TIMEOUTS: {
  aegean: number;
  consensus: number;
  parallel: number;
  reflexion: number;
  review: number;
  self-debug: number;
  self-refine: number;
  sequential: number;
};

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-schemas.ts:116

Default collaboration timeouts.

Type Declaration

aegean
readonly aegean: number;
consensus
readonly consensus: number;
parallel
readonly parallel: number;
reflexion
readonly reflexion: number;
review
readonly review: number;
self-debug
readonly self-debug: number;
self-refine
readonly self-refine: number;
sequential
readonly sequential: number;

DEFAULT_TRINITY_CONFIG

const DEFAULT_TRINITY_CONFIG: Required<TrinityConfig>;

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:182

Default TRINITY configuration.


DEFAULT_WAVE_CONFIG

const DEFAULT_WAVE_CONFIG: WaveSchedulerConfig;

Defined in: packages/nexus-agents/src/agents/wave-scheduler-types.ts:50

Default wave scheduler configuration. Matches CLAUDE.md guidelines: waves of 3-4, 2000 char output budget.


EXPERT_CAPABILITIES

const EXPERT_CAPABILITIES: Readonly<Record<AgentRole, readonly string[]>>;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:354

Expert role capabilities mapping. Maps each expert role to their core capabilities.


EXPERT_DEFAULT_CAPABILITIES

const EXPERT_DEFAULT_CAPABILITIES: Record<AgentRole, readonly AgentCapability[]>;

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:341

Default capabilities for each expert role.


EXPERT_DEFAULT_TEMPERATURES

const EXPERT_DEFAULT_TEMPERATURES: Record<ExpertDomain, number>;

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:330

Default temperatures for each expert domain.


EXPERT_TYPE_TO_ROLE

const EXPERT_TYPE_TO_ROLE: Readonly<Record<BuiltInExpertType, AgentRole>>;

Defined in: packages/nexus-agents/src/agents/experts/expert-config.ts:589

Maps built-in expert types to their AgentRole.


ExpertAssignmentSchema

const ExpertAssignmentSchema: ZodObject<{
  confidence: ZodNumber;
  expertRole: ZodEnum<{
     architecture_expert: "architecture_expert";
     code_expert: "code_expert";
     custom: "custom";
     devops_expert: "devops_expert";
     documentation_expert: "documentation_expert";
     infrastructure_expert: "infrastructure_expert";
     orchestrator: "orchestrator";
     pm_expert: "pm_expert";
     research_expert: "research_expert";
     security_expert: "security_expert";
     testing_expert: "testing_expert";
     ux_expert: "ux_expert";
  }>;
  ictmConfig: ZodOptional<ZodObject<{
     context: ZodObject<{
        includeHistory: ZodBoolean;
        maxTokens: ZodNumber;
        pruneStrategy: ZodEnum<{
           hybrid: "hybrid";
           importance: "importance";
           recency: "recency";
        }>;
        relevanceThreshold: ZodNumber;
     }, $strip>;
     instructions: ZodString;
     metadata: ZodOptional<ZodRecord<ZodString, ZodUnknown>>;
     model: ZodObject<{
        maxTokens: ZodOptional<ZodNumber>;
        modelId: ZodOptional<ZodString>;
        provider: ZodOptional<ZodString>;
        reasoning: ZodOptional<ZodEnum<{
           extended: "extended";
           minimal: "minimal";
           standard: "standard";
        }>>;
        temperature: ZodOptional<ZodNumber>;
     }, $strip>;
     tools: ZodObject<{
        capabilities: ZodArray<ZodString>;
        restrictions: ZodOptional<ZodArray<ZodString>>;
     }, $strip>;
  }, $strip>>;
  selectionReason: ZodString;
  subtaskId: ZodString;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:276

Zod schema for ExpertAssignment.


ExpertCollaborationPattern

const ExpertCollaborationPattern: {
  PAIR: "pair";
  PARALLEL: "parallel";
  REVIEW_CHAIN: "review_chain";
  SEQUENTIAL: "sequential";
};

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:21

Collaboration patterns for multi-expert tasks.

Type Declaration

PAIR
readonly PAIR: "pair" = 'pair';
PARALLEL
readonly PARALLEL: "parallel" = 'parallel';
REVIEW_CHAIN
readonly REVIEW_CHAIN: "review_chain" = 'review_chain';
SEQUENTIAL
readonly SEQUENTIAL: "sequential" = 'sequential';

ExpertConfigSchema

const ExpertConfigSchema: ZodObject<{
  capabilities: ZodArray<ZodEnum<{
     code_generation: "code_generation";
     code_review: "code_review";
     collaboration: "collaboration";
     delegation: "delegation";
     research: "research";
     task_execution: "task_execution";
     tool_use: "tool_use";
  }>>;
  id: ZodString;
  metadata: ZodOptional<ZodRecord<ZodString, ZodUnknown>>;
  modelPreference: ZodOptional<ZodObject<{
     maxTokens: ZodOptional<ZodNumber>;
     modelId: ZodOptional<ZodString>;
     provider: ZodOptional<ZodString>;
     temperature: ZodOptional<ZodNumber>;
  }, $strip>>;
  name: ZodString;
  role: ZodEnum<{
     architecture_expert: "architecture_expert";
     code_expert: "code_expert";
     custom: "custom";
     data_visualization_expert: "data_visualization_expert";
     devops_expert: "devops_expert";
     documentation_expert: "documentation_expert";
     infrastructure_expert: "infrastructure_expert";
     orchestrator: "orchestrator";
     pm_expert: "pm_expert";
     qa_expert: "qa_expert";
     research_expert: "research_expert";
     security_expert: "security_expert";
     testing_expert: "testing_expert";
     ux_expert: "ux_expert";
  }>;
  systemPrompt: ZodString;
  toolRestrictions: ZodOptional<ZodObject<{
     allowedTools: ZodOptional<ZodArray<ZodString>>;
     deniedTools: ZodOptional<ZodArray<ZodString>>;
  }, $strip>>;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/experts/expert-config.ts:147

Agents exports - Agent framework, Orchestrator, Experts Split from index.ts for file size compliance (Issue #285)


ExpertDomainSchema

const ExpertDomainSchema: ZodEnum<{
  architecture: "architecture";
  code: "code";
  documentation: "documentation";
  security: "security";
  testing: "testing";
}>;

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:241

Expert domain schema.


ExpertFactory

const ExpertFactory: {
  create: (config, options?) => Result<Expert, FactoryError>;
  createAllBuiltIn: (options?) => Result<Expert[], FactoryError>;
  createBuiltIn: (type, options?) => Result<Expert, FactoryError>;
  createFromICTM: (ictm, subtaskId, options?) => Result<Expert, FactoryError>;
  createMany: (configs, options?) => Result<Expert[], FactoryError>;
  getBuiltInConfig: (type) => Result<ExpertConfig, FactoryError>;
  validate: (config) => Result<ExpertConfig, FactoryError>;
};

Defined in: packages/nexus-agents/src/agents/experts/expert-factory.ts:429

Factory namespace for creating expert agents. Provides static methods for backward compatibility.

Type Declaration

create
readonly create: (config, options?) => Result<Expert, FactoryError> = createExpert;

Create an expert agent from a configuration object.

Parameters
config

ExpertConfig

Expert configuration

options?

CreateExpertOptions

Creation options including adapter

Returns

Result<Expert, FactoryError>

Result with Expert or FactoryError

Example
const config: ExpertConfig = {
  id: 'my-expert',
  name: 'My Expert',
  role: 'code_expert',
  capabilities: ['task_execution'],
  systemPrompt: 'You are a code review expert.',
};
const result = createExpert(config, { adapter: myAdapter });
if (result.ok) {
  const expert = result.value;
}
createAllBuiltIn
readonly createAllBuiltIn: (options?) => Result<Expert[], FactoryError> = createAllBuiltInExperts;

Create all built-in experts.

Parameters
options?

CreateExpertOptions

Creation options applied to all experts

Returns

Result<Expert[], FactoryError>

Result with array of all built-in Experts

createBuiltIn
readonly createBuiltIn: (type, options?) => Result<Expert, FactoryError> = createBuiltInExpert;

Create a built-in expert by type.

Built-in types include: ‘code’, ‘architecture’, ‘security’, ‘documentation’, and ‘testing’.

Parameters
type

BuiltInExpertType

Built-in expert type

options?

CreateExpertOptions

Creation options including adapter

Returns

Result<Expert, FactoryError>

Result with Expert or FactoryError

Example
const result = createBuiltInExpert('security', { adapter: myAdapter });
if (result.ok) {
  const securityExpert = result.value;
}
createFromICTM
createFromICTM: (ictm, subtaskId, options?) => Result<Expert, FactoryError>;

Create an expert agent from an ICTM configuration (Issue #756).

Bridges the ICTM pattern to the existing expert factory by converting the ICTM config to an ExpertConfig and delegating to createExpert().

Parameters
ictm

ICTMConfig

ICTM configuration with instructions, context, tools, model

subtaskId

string

Subtask identifier used for naming

options?

CreateExpertOptions

Creation options including adapter

Returns

Result<Expert, FactoryError>

Result with Expert or FactoryError

createMany
readonly createMany: (configs, options?) => Result<Expert[], FactoryError> = createManyExperts;

Create multiple experts from configurations.

Parameters
configs

ExpertConfig[]

Array of expert configurations

options?

CreateExpertOptions

Creation options applied to all experts

Returns

Result<Expert[], FactoryError>

Result with array of Experts or first FactoryError

getBuiltInConfig
readonly getBuiltInConfig: (type) => Result<ExpertConfig, FactoryError> = getBuiltInExpertConfig;

Get the configuration for a built-in expert type.

Parameters
type

BuiltInExpertType

Built-in expert type

Returns

Result<ExpertConfig, FactoryError>

Result with config or FactoryError if type invalid

validate
readonly validate: (config) => Result<ExpertConfig, FactoryError> = validateExpertConfigStrict;

Validate a configuration without creating an expert.

Parameters
config

unknown

Configuration to validate

Returns

Result<ExpertConfig, FactoryError>

Result with validated config or FactoryError


ExpertMatchSchema

const ExpertMatchSchema: ZodObject<{
  expertId: ZodString;
  matchedCapabilities: ZodArray<ZodString>;
  reasoning: ZodString;
  score: ZodNumber;
  scoreBreakdown: ZodObject<{
     capabilityScore: ZodNumber;
     domainScore: ZodNumber;
     finalScore: ZodNumber;
     weightScore: ZodNumber;
  }, $strip>;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:128

Agents exports - Agent framework, Orchestrator, Experts Split from index.ts for file size compliance (Issue #285)


ExpertOptionsSchema

const ExpertOptionsSchema: ZodObject<{
  additionalCapabilities: ZodOptional<ZodArray<ZodString>>;
  enableHeuristics: ZodOptional<ZodBoolean>;
  maxTokens: ZodOptional<ZodNumber>;
  systemPromptOverride: ZodOptional<ZodString>;
  temperature: ZodOptional<ZodNumber>;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:252

Expert options schema.


ExpertOutputSchema

const ExpertOutputSchema: ZodObject<{
  confidence: ZodNumber;
  content: ZodString;
  recommendations: ZodOptional<ZodArray<ZodString>>;
  structuredData: ZodOptional<ZodRecord<ZodString, ZodUnknown>>;
  warnings: ZodOptional<ZodArray<ZodString>>;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:263

Expert output schema.


ExpertParticipationSchema

const ExpertParticipationSchema: ZodObject<{
  expertId: ZodString;
  joinedAt: ZodISODateTime;
  retryCount: ZodNumber;
  role: ZodEnum<{
     architecture_expert: "architecture_expert";
     code_expert: "code_expert";
     custom: "custom";
     documentation_expert: "documentation_expert";
     orchestrator: "orchestrator";
     security_expert: "security_expert";
     testing_expert: "testing_expert";
  }>;
  status: ZodEnum<{
     failed: "failed";
     pending: "pending";
     reviewing: "reviewing";
     submitted: "submitted";
     voted: "voted";
     working: "working";
  }>;
  submittedAt: ZodOptional<ZodISODateTime>;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-schemas.ts:72

Zod schema for ExpertParticipation.


ExplorationEventSchema

const ExplorationEventSchema: ZodObject<{
  details: ZodRecord<ZodString, ZodUnknown>;
  eventType: ZodEnum<{
     conclusion_reached: "conclusion_reached";
     cross_tree_share: "cross_tree_share";
     forest_completed: "forest_completed";
     forest_converging: "forest_converging";
     node_activated: "node_activated";
     node_completed: "node_completed";
     node_created: "node_created";
     node_deactivated: "node_deactivated";
     node_pruned: "node_pruned";
     path_scored: "path_scored";
     tree_completed: "tree_completed";
     tree_created: "tree_created";
  }>;
  nodeId: ZodOptional<ZodString>;
  timestamp: ZodNumber;
  treeId: ZodOptional<ZodString>;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-result-types.ts:140

Schema for ExplorationEvent validation.


ExplorationEventTypeSchema

const ExplorationEventTypeSchema: ZodEnum<{
  conclusion_reached: "conclusion_reached";
  cross_tree_share: "cross_tree_share";
  forest_completed: "forest_completed";
  forest_converging: "forest_converging";
  node_activated: "node_activated";
  node_completed: "node_completed";
  node_created: "node_created";
  node_deactivated: "node_deactivated";
  node_pruned: "node_pruned";
  path_scored: "path_scored";
  tree_completed: "tree_completed";
  tree_created: "tree_created";
}>;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-result-types.ts:106

Schema for ExplorationEventType validation.


FailurePatternSchema

const FailurePatternSchema: ZodObject<{
  avgFailureScore: ZodNumber;
  occurrences: ZodNumber;
  pattern: ZodString;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-types.ts:98

Schema for FailurePattern validation.


ForestConfigSchema

const ForestConfigSchema: ZodObject<{
  activationBudget: ZodDefault<ZodNumber>;
  activationStrategy: ZodDefault<ZodEnum<{
     adaptive: "adaptive";
     diverse: "diverse";
     greedy: "greedy";
     ucb: "ucb";
  }>>;
  confidenceThreshold: ZodDefault<ZodNumber>;
  crossTreeStrategy: ZodDefault<ZodEnum<{
     conclusions: "conclusions";
     full: "full";
     insights: "insights";
     none: "none";
  }>>;
  earlyTerminationThreshold: ZodDefault<ZodNumber>;
  enableCrossTreeSharing: ZodDefault<ZodBoolean>;
  enableEarlyTermination: ZodDefault<ZodBoolean>;
  enableParallelExploration: ZodDefault<ZodBoolean>;
  explorationConstant: ZodDefault<ZodNumber>;
  maxDepth: ZodDefault<ZodNumber>;
  maxExplorationTimeMs: ZodDefault<ZodNumber>;
  maxNodesPerTree: ZodDefault<ZodNumber>;
  maxTokensPerTree: ZodDefault<ZodNumber>;
  maxTrees: ZodDefault<ZodNumber>;
  minScoreThreshold: ZodDefault<ZodNumber>;
  nodeTimeoutMs: ZodDefault<ZodNumber>;
  parallelThreads: ZodDefault<ZodNumber>;
  pruningStrategy: ZodDefault<ZodEnum<{
     combined: "combined";
     depth: "depth";
     none: "none";
     score: "score";
  }>>;
  seed: ZodDefault<ZodNullable<ZodNumber>>;
  sparsityRatio: ZodDefault<ZodNumber>;
  temperature: ZodDefault<ZodNumber>;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-config-types.ts:122

Schema for ForestConfig validation.


ForestPruningStrategySchema

const ForestPruningStrategySchema: ZodEnum<{
  combined: "combined";
  depth: "depth";
  none: "none";
  score: "score";
}>;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-config-types.ts:60

Schema for ForestPruningStrategy validation.


ForestResultSchema

const ForestResultSchema: ZodObject<{
  bestSolution: ZodNullable<ZodObject<{
     combinedScore: ZodNumber;
     conclusionNode: ZodObject<{
        activationScore: ZodNumber;
        children: ZodArray<ZodString>;
        confidence: ZodNumber;
        content: ZodString;
        createdAt: ZodNumber;
        depth: ZodNumber;
        estimatedValue: ZodNumber;
        id: ZodString;
        isActive: ZodBoolean;
        metadata: ZodObject<{
           crossReferences: ZodOptional<ZodArray<...>>;
           custom: ZodOptional<ZodRecord<..., ...>>;
           generationTimeMs: ZodOptional<ZodNumber>;
           source: ZodOptional<ZodString>;
           tokensUsed: ZodOptional<ZodNumber>;
        }, $strip>;
        parentId: ZodNullable<ZodString>;
        qualityScore: ZodNumber;
        state: ZodEnum<{
           active: "active";
           completed: "completed";
           error: "error";
           pending: "pending";
           pruned: "pruned";
        }>;
        stepType: ZodEnum<{
           conclusion: "conclusion";
           decomposition: "decomposition";
           hypothesis: "hypothesis";
           inference: "inference";
           synthesis: "synthesis";
           verification: "verification";
        }>;
        treeId: ZodString;
        updatedAt: ZodNumber;
     }, $strip>;
     confidence: ZodNumber;
     path: ZodArray<ZodString>;
     qualityScore: ZodNumber;
     treeId: ZodString;
  }, $strip>>;
  durationMs: ZodNumber;
  explorationHistory: ZodOptional<ZodArray<ZodObject<{
     details: ZodRecord<ZodString, ZodUnknown>;
     eventType: ZodEnum<{
        conclusion_reached: "conclusion_reached";
        cross_tree_share: "cross_tree_share";
        forest_completed: "forest_completed";
        forest_converging: "forest_converging";
        node_activated: "node_activated";
        node_completed: "node_completed";
        node_created: "node_created";
        node_deactivated: "node_deactivated";
        node_pruned: "node_pruned";
        path_scored: "path_scored";
        tree_completed: "tree_completed";
        tree_created: "tree_created";
     }>;
     nodeId: ZodOptional<ZodString>;
     timestamp: ZodNumber;
     treeId: ZodOptional<ZodString>;
  }, $strip>>>;
  finalState: ZodEnum<{
     completed: "completed";
     converging: "converging";
     exploring: "exploring";
     initializing: "initializing";
     timeout: "timeout";
  }>;
  forestId: ZodString;
  problem: ZodString;
  statistics: ZodObject<{
     activationRatio: ZodNumber;
     activeTrees: ZodNumber;
     avgTreeScore: ZodNumber;
     bestPathScore: ZodNumber;
     maxDepth: ZodNumber;
     totalActiveNodes: ZodNumber;
     totalExplorationTimeMs: ZodNumber;
     totalNodes: ZodNumber;
     totalTokensUsed: ZodNumber;
     totalTrees: ZodNumber;
  }, $strip>;
  terminationReason: ZodEnum<{
     convergence: "convergence";
     error: "error";
     max_depth: "max_depth";
     max_time: "max_time";
     max_tokens: "max_tokens";
     no_progress: "no_progress";
     solution_found: "solution_found";
  }>;
  topPaths: ZodArray<ZodObject<{
     breakdown: ZodObject<{
        coherenceScore: ZodNumber;
        conclusionBonus: ZodNumber;
        confidenceScore: ZodNumber;
        depthFactor: ZodNumber;
        qualityScore: ZodNumber;
     }, $strip>;
     length: ZodNumber;
     path: ZodArray<ZodString>;
     reachesConclusion: ZodBoolean;
     score: ZodNumber;
     targetNodeId: ZodString;
     treeId: ZodString;
  }, $strip>>;
  totalTokensUsed: ZodNumber;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-result-types.ts:187

Schema for ForestResult validation (partial).


ForestStateSchema

const ForestStateSchema: ZodEnum<{
  completed: "completed";
  converging: "converging";
  exploring: "exploring";
  initializing: "initializing";
  timeout: "timeout";
}>;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-state-types.ts:33

Schema for ForestState validation.


ForestStatisticsSchema

const ForestStatisticsSchema: ZodObject<{
  activationRatio: ZodNumber;
  activeTrees: ZodNumber;
  avgTreeScore: ZodNumber;
  bestPathScore: ZodNumber;
  maxDepth: ZodNumber;
  totalActiveNodes: ZodNumber;
  totalExplorationTimeMs: ZodNumber;
  totalNodes: ZodNumber;
  totalTokensUsed: ZodNumber;
  totalTrees: ZodNumber;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-state-types.ts:74

Schema for ForestStatistics validation.


GeneratedTestSchema

const GeneratedTestSchema: ZodObject<{
  code: ZodString;
  name: ZodString;
  scenarios: ZodArray<ZodString>;
  target: ZodString;
  type: ZodEnum<{
     e2e: "e2e";
     integration: "integration";
     unit: "unit";
  }>;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:308

Generated test schema.


MIN_EXPERTS_FOR_PATTERN

const MIN_EXPERTS_FOR_PATTERN: {
  aegean: 3;
  consensus: 3;
  parallel: 2;
  reflexion: 1;
  review: 2;
  self-debug: 1;
  self-refine: 1;
  sequential: 1;
};

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-schemas.ts:135

Minimum number of experts for each pattern.

Type Declaration

aegean
readonly aegean: 3 = 3;
consensus
readonly consensus: 3 = 3;
parallel
readonly parallel: 2 = 2;
reflexion
readonly reflexion: 1 = 1;
review
readonly review: 2 = 2;
self-debug
readonly self-debug: 1 = 1;
self-refine
readonly self-refine: 1 = 1;
sequential
readonly sequential: 1 = 1;

ModelPreferenceSchema

const ModelPreferenceSchema: ZodObject<{
  maxTokens: ZodOptional<ZodNumber>;
  modelId: ZodOptional<ZodString>;
  provider: ZodOptional<ZodString>;
  temperature: ZodOptional<ZodNumber>;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/experts/expert-config.ts:87

Zod schema for ModelPreference.


NodeStateSchema

const NodeStateSchema: ZodEnum<{
  active: "active";
  completed: "completed";
  error: "error";
  pending: "pending";
  pruned: "pruned";
}>;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-node-types.ts:50

Schema for NodeState validation.


OrchestratorOptionsSchema

const OrchestratorOptionsSchema: ZodObject<{
  decompositionThreshold: ZodOptional<ZodNumber>;
  enableParallelHints: ZodOptional<ZodBoolean>;
  expertWeights: ZodOptional<ZodRecord<ZodString, ZodNumber>>;
  maxSubtasks: ZodOptional<ZodNumber>;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:343

Zod schema for OrchestratorOptions.


PathScoreBreakdownSchema

const PathScoreBreakdownSchema: ZodObject<{
  coherenceScore: ZodNumber;
  conclusionBonus: ZodNumber;
  confidenceScore: ZodNumber;
  depthFactor: ZodNumber;
  qualityScore: ZodNumber;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:56

Schema for PathScoreBreakdown validation.


PathScoreSchema

const PathScoreSchema: ZodObject<{
  breakdown: ZodObject<{
     coherenceScore: ZodNumber;
     conclusionBonus: ZodNumber;
     confidenceScore: ZodNumber;
     depthFactor: ZodNumber;
     qualityScore: ZodNumber;
  }, $strip>;
  length: ZodNumber;
  path: ZodArray<ZodString>;
  reachesConclusion: ZodBoolean;
  score: ZodNumber;
  targetNodeId: ZodString;
  treeId: ZodString;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:87

Schema for PathScore validation.


PruningStrategy

const PruningStrategy: {
  HIERARCHICAL: "hierarchical";
  LOWEST_PRIORITY: "lowest_priority";
  OLDEST_FIRST: "oldest_first";
  PRIORITY_WEIGHTED_AGE: "priority_weighted_age";
  SEMANTIC: "semantic";
  SLIDING_WINDOW: "sliding_window";
  SUMMARIZE: "summarize";
};

Defined in: packages/nexus-agents/src/agents/context-pruner.ts:41

Strategy for pruning context when budget is exceeded.

Type Declaration

HIERARCHICAL
readonly HIERARCHICAL: "hierarchical" = 'hierarchical';
LOWEST_PRIORITY
readonly LOWEST_PRIORITY: "lowest_priority" = 'lowest_priority';
OLDEST_FIRST
readonly OLDEST_FIRST: "oldest_first" = 'oldest_first';
PRIORITY_WEIGHTED_AGE
readonly PRIORITY_WEIGHTED_AGE: "priority_weighted_age" = 'priority_weighted_age';
SEMANTIC
readonly SEMANTIC: "semantic" = 'semantic';
SLIDING_WINDOW
readonly SLIDING_WINDOW: "sliding_window" = 'sliding_window';
SUMMARIZE
readonly SUMMARIZE: "summarize" = 'summarize';

ReasoningNodeMetadataSchema

const ReasoningNodeMetadataSchema: ZodObject<{
  crossReferences: ZodOptional<ZodArray<ZodString>>;
  custom: ZodOptional<ZodRecord<ZodString, ZodUnknown>>;
  generationTimeMs: ZodOptional<ZodNumber>;
  source: ZodOptional<ZodString>;
  tokensUsed: ZodOptional<ZodNumber>;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-node-types.ts:105

Schema for ReasoningNodeMetadata validation.


ReasoningNodeSchema

const ReasoningNodeSchema: ZodObject<{
  activationScore: ZodNumber;
  children: ZodArray<ZodString>;
  confidence: ZodNumber;
  content: ZodString;
  createdAt: ZodNumber;
  depth: ZodNumber;
  estimatedValue: ZodNumber;
  id: ZodString;
  isActive: ZodBoolean;
  metadata: ZodObject<{
     crossReferences: ZodOptional<ZodArray<ZodString>>;
     custom: ZodOptional<ZodRecord<ZodString, ZodUnknown>>;
     generationTimeMs: ZodOptional<ZodNumber>;
     source: ZodOptional<ZodString>;
     tokensUsed: ZodOptional<ZodNumber>;
  }, $strip>;
  parentId: ZodNullable<ZodString>;
  qualityScore: ZodNumber;
  state: ZodEnum<{
     active: "active";
     completed: "completed";
     error: "error";
     pending: "pending";
     pruned: "pruned";
  }>;
  stepType: ZodEnum<{
     conclusion: "conclusion";
     decomposition: "decomposition";
     hypothesis: "hypothesis";
     inference: "inference";
     synthesis: "synthesis";
     verification: "verification";
  }>;
  treeId: ZodString;
  updatedAt: ZodNumber;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-node-types.ts:159

Schema for ReasoningNode validation.


ReasoningStepTypeSchema

const ReasoningStepTypeSchema: ZodEnum<{
  conclusion: "conclusion";
  decomposition: "decomposition";
  hypothesis: "hypothesis";
  inference: "inference";
  synthesis: "synthesis";
  verification: "verification";
}>;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-node-types.ts:73

Schema for ReasoningStepType validation.


ReasoningTreeSchema

const ReasoningTreeSchema: ZodObject<{
  createdAt: ZodNumber;
  explorationPriority: ZodNumber;
  forestId: ZodString;
  hypothesis: ZodString;
  id: ZodString;
  overallScore: ZodNumber;
  rootId: ZodString;
  state: ZodEnum<{
     abandoned: "abandoned";
     completed: "completed";
     growing: "growing";
     paused: "paused";
  }>;
  statistics: ZodObject<{
     activeNodes: ZodNumber;
     avgBranchingFactor: ZodNumber;
     avgConfidence: ZodNumber;
     avgQualityScore: ZodNumber;
     conclusionCount: ZodNumber;
     maxDepth: ZodNumber;
     totalNodes: ZodNumber;
     totalTokensUsed: ZodNumber;
  }, $strip>;
  updatedAt: ZodNumber;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:205

Schema for ReasoningTree validation (partial, excludes Map for JSON).


ReviewResponseMessageSchema

const ReviewResponseMessageSchema: ZodObject<{
  approved: ZodBoolean;
  feedback: ZodString;
  requesterId: ZodString;
  reviewerId: ZodString;
  severity: ZodOptional<ZodEnum<{
     critical: "critical";
     major: "major";
     minor: "minor";
     none: "none";
  }>>;
  suggestions: ZodOptional<ZodArray<ZodString>>;
  type: ZodLiteral<"review_response">;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-schemas.ts:103

Zod schema for ReviewResponseMessage.


ScoreBreakdownSchema

const ScoreBreakdownSchema: ZodObject<{
  capabilityScore: ZodNumber;
  domainScore: ZodNumber;
  finalScore: ZodNumber;
  weightScore: ZodNumber;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:121

Agents exports - Agent framework, Orchestrator, Experts Split from index.ts for file size compliance (Issue #285)


SelectionOptionsSchema

const SelectionOptionsSchema: ZodObject<{
  capabilityWeights: ZodOptional<ZodRecord<ZodString, ZodNumber>>;
  excludeExperts: ZodOptional<ZodArray<ZodString>>;
  forceCollaboration: ZodOptional<ZodBoolean>;
  maxAlternatives: ZodOptional<ZodNumber>;
  minScore: ZodOptional<ZodNumber>;
  preferredDomains: ZodOptional<ZodArray<ZodEnum<{
     architecture: "architecture";
     code: "code";
     documentation: "documentation";
     general: "general";
     security: "security";
     testing: "testing";
  }>>>;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:144

Agents exports - Agent framework, Orchestrator, Experts Split from index.ts for file size compliance (Issue #285)


SelectionResultSchema

const SelectionResultSchema: ZodObject<{
  alternatives: ZodArray<ZodObject<{
     expertId: ZodString;
     matchedCapabilities: ZodArray<ZodString>;
     reasoning: ZodString;
     score: ZodNumber;
     scoreBreakdown: ZodObject<{
        capabilityScore: ZodNumber;
        domainScore: ZodNumber;
        finalScore: ZodNumber;
        weightScore: ZodNumber;
     }, $strip>;
  }, $strip>>;
  confidence: ZodNumber;
  primary: ZodObject<{
     expertId: ZodString;
     matchedCapabilities: ZodArray<ZodString>;
     reasoning: ZodString;
     score: ZodNumber;
     scoreBreakdown: ZodObject<{
        capabilityScore: ZodNumber;
        domainScore: ZodNumber;
        finalScore: ZodNumber;
        weightScore: ZodNumber;
     }, $strip>;
  }, $strip>;
  requiresCollaboration: ZodBoolean;
  suggestedPattern: ZodOptional<ZodEnum<{
     pair: "pair";
     parallel: "parallel";
     review_chain: "review_chain";
     sequential: "sequential";
  }>>;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/experts/expert-selector-types.ts:136

Agents exports - Agent framework, Orchestrator, Experts Split from index.ts for file size compliance (Issue #285)


SessionStatusSchema

const SessionStatusSchema: ZodEnum<{
  awaiting_review: "awaiting_review";
  completed: "completed";
  failed: "failed";
  finalizing: "finalizing";
  in_progress: "in_progress";
  pending: "pending";
  timed_out: "timed_out";
  voting: "voting";
}>;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-schemas.ts:26

Zod schema for SessionStatus.


SharedConclusionSchema

const SharedConclusionSchema: ZodObject<{
  confidence: ZodNumber;
  content: ZodString;
  qualityScore: ZodNumber;
  sourceNodeId: ZodString;
  sourceTreeId: ZodString;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-types.ts:51

Schema for SharedConclusion validation.


SharedInsightSchema

const SharedInsightSchema: ZodObject<{
  content: ZodString;
  relevance: ZodNumber;
  sourceNodeId: ZodString;
  sourceTreeId: ZodString;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-types.ts:76

Schema for SharedInsight validation.


SubtaskPrioritySchema

const SubtaskPrioritySchema: ZodEnum<{
  critical: "critical";
  high: "high";
  low: "low";
  medium: "medium";
}>;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:192

Zod schema for SubtaskPriority.


SubTaskSchema

const SubTaskSchema: ZodObject<{
  assignedRole: ZodOptional<ZodEnum<{
     architecture_expert: "architecture_expert";
     code_expert: "code_expert";
     custom: "custom";
     devops_expert: "devops_expert";
     documentation_expert: "documentation_expert";
     infrastructure_expert: "infrastructure_expert";
     orchestrator: "orchestrator";
     pm_expert: "pm_expert";
     research_expert: "research_expert";
     security_expert: "security_expert";
     testing_expert: "testing_expert";
     ux_expert: "ux_expert";
  }>>;
  complexity: ZodNumber;
  dependencies: ZodArray<ZodString>;
  description: ZodString;
  expectedOutput: ZodString;
  id: ZodString;
  parentTaskId: ZodString;
  priority: ZodEnum<{
     critical: "critical";
     high: "high";
     low: "low";
     medium: "medium";
  }>;
  requiredCapabilities: ZodArray<ZodString>;
  status: ZodEnum<{
     assigned: "assigned";
     completed: "completed";
     failed: "failed";
     in_progress: "in_progress";
     pending: "pending";
  }>;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:208

Zod schema for SubTask.


SubtaskStatusSchema

const SubtaskStatusSchema: ZodEnum<{
  assigned: "assigned";
  completed: "completed";
  failed: "failed";
  in_progress: "in_progress";
  pending: "pending";
}>;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:197

Zod schema for SubtaskStatus.


SynthesizedResultSchema

const SynthesizedResultSchema: ZodObject<{
  collaborationMetadata: ZodOptional<ZodObject<{
     agreementLevel: ZodNumber;
     participantCount: ZodNumber;
     pattern: ZodString;
     sessionId: ZodString;
  }, $strip>>;
  combinedOutput: ZodString;
  conflicts: ZodArray<ZodObject<{
     description: ZodString;
     resolution: ZodString;
     subtaskId1: ZodString;
     subtaskId2: ZodString;
  }, $strip>>;
  qualityScore: ZodNumber;
  recommendations: ZodArray<ZodString>;
  resultSummaries: ZodArray<ZodObject<{
     contributions: ZodArray<ZodString>;
     quality: ZodNumber;
     subtaskId: ZodString;
     summary: ZodString;
  }, $strip>>;
  summary: ZodString;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:330

Zod schema for SynthesizedResult.


TASK_TYPE_EXPERTS

const TASK_TYPE_EXPERTS: Readonly<Record<string, AgentRole>>;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:379

Task type to expert role mapping. Maps common task types to their primary expert roles.


TaskAnalysisSchema

const TaskAnalysisSchema: ZodObject<{
  approach: ZodString;
  commitment: ZodOptional<ZodObject<{
     approach: ZodString;
     constraints: ZodArray<ZodString>;
     differentiation: ZodString;
     purpose: ZodString;
  }, $strip>>;
  complexity: ZodPipe<ZodUnion<readonly [ZodNumber, ZodString]>, ZodTransform<number, string | number>>;
  estimatedEffort: ZodPipe<ZodUnion<readonly [ZodNumber, ZodString]>, ZodTransform<number, string | number>>;
  needsDecomposition: ZodPipe<ZodUnion<readonly [ZodBoolean, ZodString]>, ZodTransform<boolean, string | boolean>>;
  requirements: ZodArray<ZodString>;
  risks: ZodArray<ZodString>;
  taskId: ZodString;
  taskType: ZodString;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/tech-lead-types.ts:244

Zod schema for TaskAnalysis. Uses coercion and transforms for numeric fields because LLMs may return numbers as strings or descriptive words (Issue #663).


TaskSchema

const TaskSchema: ZodObject<{
  constraints: ZodOptional<ZodObject<{
     maxDuration: ZodOptional<ZodNumber>;
     maxTokens: ZodOptional<ZodNumber>;
  }, $strip>>;
  context: ZodObject<{
     files: ZodOptional<ZodArray<ZodString>>;
     history: ZodOptional<ZodArray<ZodObject<{
        content: ZodString;
        role: ZodEnum<{
           assistant: "assistant";
           system: "system";
           user: "user";
        }>;
        timestamp: ZodString;
     }, $strip>>>;
     metadata: ZodOptional<ZodRecord<ZodString, ZodUnknown>>;
     workingDirectory: ZodOptional<ZodString>;
  }, $strip>;
  description: ZodString;
  id: ZodString;
  priority: ZodOptional<ZodNumber>;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/agent-schemas.ts:13

Zod schema for validating Task objects.


TerminationReasonSchema

const TerminationReasonSchema: ZodEnum<{
  convergence: "convergence";
  error: "error";
  max_depth: "max_depth";
  max_time: "max_time";
  max_tokens: "max_tokens";
  no_progress: "no_progress";
  solution_found: "solution_found";
}>;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-result-types.ts:38

Schema for TerminationReason validation.


TreeStateSchema

const TreeStateSchema: ZodEnum<{
  abandoned: "abandoned";
  completed: "completed";
  growing: "growing";
  paused: "paused";
}>;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:31

Schema for TreeState validation.


TreeStatisticsSchema

const TreeStatisticsSchema: ZodObject<{
  activeNodes: ZodNumber;
  avgBranchingFactor: ZodNumber;
  avgConfidence: ZodNumber;
  avgQualityScore: ZodNumber;
  conclusionCount: ZodNumber;
  maxDepth: ZodNumber;
  totalNodes: ZodNumber;
  totalTokensUsed: ZodNumber;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/reasoning/forest-tree-types.ts:153

Schema for TreeStatistics validation.


TRINITY_ROLE_MAX_TOKENS

const TRINITY_ROLE_MAX_TOKENS: Record<TrinityRole, number>;

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:91

Default max tokens for TRINITY roles.


TRINITY_ROLE_PROMPTS

const TRINITY_ROLE_PROMPTS: Record<TrinityRole, string>;

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:35

Default prompts for each TRINITY role.


TRINITY_ROLE_TEMPERATURES

const TRINITY_ROLE_TEMPERATURES: Record<TrinityRole, number>;

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:84

Default temperatures for TRINITY roles.


TrinityConfigSchema

const TrinityConfigSchema: ZodObject<{
  includeHistory: ZodOptional<ZodBoolean>;
  maxIterations: ZodOptional<ZodNumber>;
  roleConfigs: ZodOptional<ZodRecord<ZodEnum<{
     thinker: "thinker";
     verifier: "verifier";
     worker: "worker";
   }>, ZodObject<{
     maxTokens: ZodOptional<ZodNumber>;
     systemPrompt: ZodOptional<ZodString>;
     temperature: ZodOptional<ZodNumber>;
  }, $strip>>>;
  timeoutMs: ZodOptional<ZodNumber>;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:229

Schema for TrinityConfig.


TrinityPhaseSchema

const TrinityPhaseSchema: ZodEnum<{
  complete: "complete";
  thinking: "thinking";
  verifying: "verifying";
  working: "working";
}>;

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:223

Schema for TRINITY phase.


TrinityRoleSchema

const TrinityRoleSchema: ZodEnum<{
  thinker: "thinker";
  verifier: "verifier";
  worker: "worker";
}>;

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:220

Schema for TRINITY role.


TrinityStopReasonSchema

const TrinityStopReasonSchema: ZodEnum<{
  error: "error";
  max_iterations: "max_iterations";
  timeout: "timeout";
  verified: "verified";
}>;

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:246

Schema for stop reason.


VerifierVerdictSchema

const VerifierVerdictSchema: ZodEnum<{
  fail: "fail";
  pass: "pass";
}>;

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-types.ts:226

Schema for verifier verdict.


VoteDecisionSchema

const VoteDecisionSchema: ZodEnum<{
  abstain: "abstain";
  approve: "approve";
  reject: "reject";
}>;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-schemas.ts:40

Zod schema for VoteDecision.


VoteMessageSchema

const VoteMessageSchema: ZodObject<{
  conditions: ZodOptional<ZodArray<ZodString>>;
  decision: ZodEnum<{
     abstain: "abstain";
     approve: "approve";
     reject: "reject";
  }>;
  expertId: ZodString;
  reasoning: ZodString;
  type: ZodLiteral<"vote">;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-schemas.ts:92

Zod schema for VoteMessage.


VulnerabilitySchema

const VulnerabilitySchema: ZodObject<{
  cweId: ZodOptional<ZodString>;
  description: ZodString;
  id: ZodString;
  location: ZodOptional<ZodString>;
  remediation: ZodString;
  severity: ZodEnum<{
     critical: "critical";
     high: "high";
     info: "info";
     low: "low";
     medium: "medium";
  }>;
  type: ZodString;
}, $strip>;

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:279

Vulnerability schema.


VulnerabilitySeveritySchema

const VulnerabilitySeveritySchema: ZodEnum<{
  critical: "critical";
  high: "high";
  info: "info";
  low: "low";
  medium: "medium";
}>;

Defined in: packages/nexus-agents/src/agents/experts/expert-types.ts:274

Vulnerability severity schema.

Functions

aggregateResults()

function aggregateResults(input, options?): Result<AggregatedResult, AgentError>;

Defined in: packages/nexus-agents/src/agents/collaboration/result-aggregator.ts:241

Convenience function to aggregate results.

Parameters

input

AggregatorInput

options?

AggregatorOptions

Returns

Result<AggregatedResult, AgentError>


chunkByDirectory()

function chunkByDirectory(files, basePath): WorkChunk[];

Defined in: packages/nexus-agents/src/agents/wave-scheduler.ts:376

Partition a list of file paths into directory-scoped chunks. Each chunk corresponds to a top-level directory within the base path.

Parameters

files

readonly string[]

Array of file paths

basePath

string

Base path prefix to strip for grouping

Returns

WorkChunk[]

Array of work chunks grouped by top-level directory


classifyExpertFailure()

function classifyExpertFailure(
   error, 
   detector, 
   taskDescription?, 
   signal?
): FailureClassification;

Defined in: packages/nexus-agents/src/agents/experts/expert-recovery.ts:198

Classifies a single failed execution attempt as transient or permanent.

See the module header for the ordered decision procedure. signal is checked first (mandatory guard): RETRYABLE_ERROR_PATTERNS matches /aborted/i, so without this a cancelled task would otherwise be retried.

Fails CLOSED on a throwing classifier (#4303): the body reads error.message and walks .cause (via getErrorMessage/extractErrorMessage/isRetryableErrorChain) and calls detector.detect. An Error-like object with a throwing .message or .cause getter would make any of those throw. execute() runs this INSIDE withRetry’s isRetryable predicate and again in annotateExhausted, neither of which is try-guarded by withRetry (adapters/retry.ts:339-363 catches only operation()), so an escaping throw would reject the Promise<Result<…>> and break the never-throws contract execute_expert relies on. This single outer guard covers BOTH call sites: any throw during classification → permanent.

Parameters

error

unknown

detector

FailureDetector

taskDescription?

string

signal?

AbortSignal

Returns

FailureClassification


createAgenticAdapter()

function createAgenticAdapter(modelAdapter, options?): IAgenticAdapter;

Defined in: packages/nexus-agents/src/agents/agentic/factory.ts:25

Build an IAgenticAdapter for the supplied model adapter.

Stamps adapterStrategy based on the model’s providerId so downstream eval results record which path they exercised. Future provider-specialised concretes will set their own strategy.

Parameters

modelAdapter

IModelAdapter

options?

AgenticAdapterOptions = {}

Returns

IAgenticAdapter


createArchitectureExpert()

function createArchitectureExpert(options?): ArchitectureExpert;

Defined in: packages/nexus-agents/src/agents/experts/architecture-expert.ts:244

Agents exports - Agent framework, Orchestrator, Experts Split from index.ts for file size compliance (Issue #285)

Parameters

options?

Partial<BaseAgentOptions> & { expertOptions?: ArchitectureExpertOptions; }

Returns

ArchitectureExpert


createCodeExpert()

function createCodeExpert(options?): CodeExpert;

Defined in: packages/nexus-agents/src/agents/experts/code-expert.ts:221

Creates a new CodeExpert agent with the given options.

Parameters

options?

Partial<BaseAgentOptions> & { expertOptions?: CodeExpertOptions; }

Returns

CodeExpert


createCollaborationSession()

function createCollaborationSession(options?): CollaborationSession;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-session.ts:449

Agents exports - Agent framework, Orchestrator, Experts Split from index.ts for file size compliance (Issue #285)

Parameters

options?

CollaborationSessionOptions

Returns

CollaborationSession


createDefaultRegistry()

function createDefaultRegistry(): SelectionExpertRegistry;

Defined in: packages/nexus-agents/src/agents/experts/expert-selector.ts:114

Creates a default expert registry with built-in experts.

Returns

SelectionExpertRegistry


createDocumentationExpert()

function createDocumentationExpert(options?): DocumentationExpert;

Defined in: packages/nexus-agents/src/agents/experts/documentation-expert.ts:312

Agents exports - Agent framework, Orchestrator, Experts Split from index.ts for file size compliance (Issue #285)

Parameters

options?

Partial<BaseAgentOptions> & { expertOptions?: DocumentationExpertOptions; }

Returns

DocumentationExpert


createOrchestrator()

function createOrchestrator(options?): Orchestrator;

Defined in: packages/nexus-agents/src/agents/tech-lead.ts:597

Creates a new Orchestrator agent with the given options. This is the preferred factory function for creating coordination agents.

Parameters

options?

Partial<BaseAgentOptions> & { orchestratorOptions?: OrchestratorOptions; }

Agent configuration options

Returns

Orchestrator

Orchestrator agent instance

Example

const orchestrator = createOrchestrator({
  orchestratorOptions: { maxSubtasks: 5 }
});
const result = await orchestrator.execute(task);

createProtocolFactory()

function createProtocolFactory(options?): ProtocolFactory;

Defined in: packages/nexus-agents/src/agents/collaboration/collaboration-protocol.ts:312

Creates a protocol factory.

Parameters

options?

ProtocolOptions

Returns

ProtocolFactory


createResultAggregator()

function createResultAggregator(options?): ResultAggregator;

Defined in: packages/nexus-agents/src/agents/collaboration/result-aggregator.ts:234

Creates a result aggregator.

Parameters

options?

AggregatorOptions

Returns

ResultAggregator


createSecurityExpert()

function createSecurityExpert(options?): SecurityExpert;

Defined in: packages/nexus-agents/src/agents/experts/security-expert.ts:249

Agents exports - Agent framework, Orchestrator, Experts Split from index.ts for file size compliance (Issue #285)

Parameters

options?

Partial<BaseAgentOptions> & { expertOptions?: SecurityExpertOptions; }

Returns

SecurityExpert


createStateMachine()

function createStateMachine(options?): AgentStateMachine;

Defined in: packages/nexus-agents/src/agents/state-machine.ts:333

Creates a new agent state machine.

Parameters

options?

StateMachineOptions

State machine options

Returns

AgentStateMachine

A new AgentStateMachine instance


createTestingExpert()

function createTestingExpert(options?): TestingExpert;

Defined in: packages/nexus-agents/src/agents/experts/testing-expert.ts:347

Agents exports - Agent framework, Orchestrator, Experts Split from index.ts for file size compliance (Issue #285)

Parameters

options?

Partial<BaseAgentOptions> & { expertOptions?: TestingExpertOptions; }

Returns

TestingExpert


createTrinityCoordinator()

function createTrinityCoordinator(config?): TrinityCoordinator;

Defined in: packages/nexus-agents/src/agents/collaboration/trinity-coordinator.ts:447

Creates a TRINITY coordinator instance.

Parameters

config?

TrinityConfig

Returns

TrinityCoordinator


createWaveScheduler()

function createWaveScheduler(config?, logger?): WaveScheduler;

Defined in: packages/nexus-agents/src/agents/wave-scheduler.ts:408

Create a new WaveScheduler instance with the given configuration.

Parameters

config?

Partial<WaveSchedulerConfig> = {}

logger?

ILogger

Returns

WaveScheduler


getExpertRegistry()

function getExpertRegistry(): ExpertRegistry;

Defined in: packages/nexus-agents/src/agents/experts/expert-registry.ts:402

Get the global expert registry instance.

Returns

ExpertRegistry


quickSelect()

function quickSelect(task, options?): Result<SelectionResult, SelectionError>;

Defined in: packages/nexus-agents/src/agents/experts/expert-selector.ts:404

Quick selection using default registry. Convenience function for simple use cases. Uses a cached registry for performance optimization.

Parameters

task

Task

options?

SelectionOptions

Returns

Result<SelectionResult, SelectionError>


safeValidateExpertConfig()

function safeValidateExpertConfig(config): 
  | {
  data: ExpertConfig;
  success: true;
}
  | {
  error: ZodError;
  success: false;
};

Defined in: packages/nexus-agents/src/agents/experts/expert-config.ts:618

Safely validates an expert configuration.

Parameters

config

unknown

Configuration to validate

Returns

| { data: ExpertConfig; success: true; } | { error: ZodError; success: false; }

Safe parse result with success/error


selectExperts()

function selectExperts(
   task, 
   registry, 
   options?
): Result<SelectionResult, SelectionError>;

Defined in: packages/nexus-agents/src/agents/experts/expert-selector.ts:370

Selects the best experts for a task.

Parameters

task

Task

The task to select experts for

registry

SelectionExpertRegistry

Registry of available experts

options?

SelectionOptions

Optional selection configuration

Returns

Result<SelectionResult, SelectionError>


validateExpertConfig()

function validateExpertConfig(config): ExpertConfig;

Defined in: packages/nexus-agents/src/agents/experts/expert-config.ts:609

Validates an expert configuration.

Parameters

config

unknown

Configuration to validate

Returns

ExpertConfig

Parsed config or throws on validation error