RunContext

Properties

abortSignal?

optional abortSignal?: AbortSignal;

---

autoRetrieve?

optional autoRetrieve?: AutoRetrieveProvider;

---

bargeIn?

optional bargeIn?: AbortSignal;

---

baseInstructions?

optional baseInstructions?: Instructions;

Agent base layer (ADR 0001), set when entering a flow. baseInstructions is composed as a prefix into every node turn’s system prompt (persona / safety / grounding floor); globalTools are safe tools made model-visible in every speaking turn.

---

controlModel

controlModel: LanguageModel;

Control-path model (routing, decide, extraction) at temperature 0.

---

emit

emit: (part) => void;

Parameters

Parameter	Type
part	HarnessStreamPart

Returns

void

---

fs?

optional fs?: FileSystem;

Agent workspace filesystem (same instance as AgentConfig.workspace).

---

globalTools?

optional globalTools?: Record<string, AnyTool>;

---

hookRunner

hookRunner: HookRunner;

---

inputProcessors

inputProcessors: InputProcessor[];

---

lastRetrievalCitations?

optional lastRetrievalCitations?: SourceRef[];

Citations from the latest gather-phase retrieval on this turn.

---

limits?

optional limits?: Limits;

---

memoryService?

optional memoryService?: MemoryService;

---

model

model: LanguageModel;

---

outOfBandControl

outOfBandControl: boolean;

When true, flow reply nodes use the out-of-band control evaluator (ADR 0003 H1).

---

outputProcessors

outputProcessors: OutputProcessor[];

---

refinementPolicies

refinementPolicies: RefinementCapability[];

---

runState

runState: RunState;

---

runStore

runStore: RunStore;

---

session

session: Session;

---

skillPrompt?

optional skillPrompt?: string;

Level-1 skill metadata injected by SkillsCapability when AgentConfig.skills is set.

---

telemetry?

optional telemetry?: TelemetrySettings;

---

toolExecutor

toolExecutor: EffectToolExecutor;

---

turnInputConsumed?

optional turnInputConsumed?: boolean;

Ephemeral, per-run-invocation flag: has the current turn’s user input been consumed yet by an input-node (collect/decide)? Input-nodes extract/decide from the turn’s fresh input; once it is consumed, later nodes in the same turn pause (present prompt, await next turn) instead of acting on stale context. Reset to false on every createRunContext (i.e. every turn).

---

validationPolicies

validationPolicies: ValidationCapability[];

---

workingMemoryPrompt?

optional workingMemoryPrompt?: string;

Frozen persistent memory blocks loaded at session start (AgentMemory.workingMemory).

---

workingMemoryTools?

optional workingMemoryTools?: Record<string, AnyTool>;

Model-visible tools from working memory wiring (not in globalTools — mutating but scoped).

Methods

approve()

approve(req): Promise<{
  approved: boolean;
  by?: string;
}>;

Parameters

Parameter	Type
req	{ description?: string; title: string; }
req.description?	string
req.title	string

Returns

Promise<{ approved: boolean; by?: string; }>

---

now()

now(): Promise<number>;

Returns

Promise<number>

---

resetCallsites()

resetCallsites(): void;

Rebase durable effect callsites to 0. The runtime calls this at flow entry so a flow’s durable callsites are anchored to the flow — identical on fresh entry (after an answering turn) and on resume (where that turn does not re-run).

Returns

void

---

signal()

signal(name, opts?): Promise<unknown>;

Parameters

Parameter	Type
name	string
opts?	{ deadline?: number; meta?: Record<string, unknown>; }
opts.deadline?	number
opts.meta?	Record<string, unknown>

Returns

Promise<unknown>

---

tool()

tool(
   name,
   args,
options?): Promise<unknown>;

Parameters

Parameter	Type
name	string
args	unknown
options?	{ def?: AnyTool; toolCallId?: string; toolCtx?: ToolContext; }
options.def?	AnyTool
options.toolCallId?	string
options.toolCtx?	ToolContext

Returns

Promise<unknown>

---

uuid()

uuid(): Promise<string>;

Returns

Promise<string>