deepseek-harness/docs/core-data-structures/core.md

Core Data Structures

English | 中文

This folder catalogs the data structures of the DeepSeek Harness — what each core type represents, its literal shape, and where the full detail lives. It complements architecture.md, which describes behavior (the service map, the session/turn/step lifecycle, the event taxonomy); this page describes the vocabulary that behavior moves around.

What counts as "core"

The harness is a microkernel: a tiny core plus many plugins. Most types belong to one plugin or one capability. A handful, though, are the spine — the language the agent loop and its events traffic in on every turn, no matter which optional plugins are loaded. Those are "core".

Precisely, a data structure is core if either:

1. it flows through the agent-loop spine — the loop holds it, derives it, streams it, or logs it on every turn (a Message, a StreamChunk, a SessionEvent, the Agent handle itself), independent of which plugins are present; or
2. it is the single headline type a plugin author writes against a pipeline — ToolDefinition (what every tool is).

Everything else is documented on a sub-page, not here. The rule that draws the line: the type you write, hold, or receive is core; the machinery that types it, renders it, or persists it is a sub-page detail. So ToolDefinition is core, but the ValueSchemaSpec/ParameterSchemaSpec inference machinery that types it, the ToolCallView/ToolResultView render-intent vocabulary that renders it, and the SessionPersistence seam that stores the event log are not — they live on the sub-pages below.

Sub-page	Owns
llm-streaming.md	the StreamChunk wire protocol + adapter contract, BlockAssembler, the LlmAdapter seam
token-meter.md	immutable scalar and positional replay measurements with consumed-log revisions
scope.md	scoped registration identity, dispatch carriers, and the owned Scope context
goal.md	persisted goal identity, lifecycle snapshots, activation, change records, and round attribution
commands.md	the human-command seam: definitions, adapter discovery, direct invocation, results, and parsing views
session.md	the full SessionEventMap variant catalog, TurnTrigger/TurnEndReason, deriveMessages(), execution enclosure, and standalone events
persistence.md	the durability seam: SessionPersistence, JSONL + SQLite backends, session/flush, crash recovery, SessionHeader
settings.md	the user-settings seam: SettingsNamespace registration, layered resolution (defaults → composition base → user document), owner scopes, hot commits
credentials.md	the credential seam: CredentialRef references (never values) in configuration, per-operation resolution, UI-safe CredentialInfo, provider source layers
session-query.md	logical records, bounded exact-event reads, relationship traces, semantic filters/documents, and full-text result pages
session-title.md	durable title snapshots, source provenance, and the asynchronous provider contract
system-prompt.md	per-assembly context, tool-provider results, prompt sections, and cooperative assembly
tools.md	ToolDefinition full fields, the schema DSL, ToolExecution/ToolResult, tool-presentation UI types, and the guarded execution pipeline
user-interaction.md	the UI-backed human question/answer seam: AskUserQuestionRequest, answer/options vocabulary, provider API, error taxonomy
approval.md	the one-shot user-approval seam: ApprovalRequest, ApprovalOutcome, per-session policy, audit and answerer contracts
bash.md	the bash executor seam: BashExecRequest/Spec, BashRunResult, background BashProcess handles
subprocess.md	the subprocess seam: fully-explicit SubprocessSpawnSpec, offset-based output readers, unclassified SubprocessOutcome, and the managed DSH_* environment vocabulary
pty.md	persistent terminal ids, backend/session contracts, send readiness, bounded reads, and owner-visible snapshots
sandbox.md	per-session policy resolution and the process-confinement seam: file-effect modes, execution/provider policies, ConfinedArgv, enforcement and fail-closed errors
code-runtime.md	the code-execution seam: CodeRunRequest/Result, binding namespaces, captured logs, the CodeRunFailure taxonomy
filesystem.md	the filesystem seam: FsTarget, read/write/edit outcomes, observed-file state, FsErrorCode
lsp.md	the LSP navigation seam: LspQueryRequest/Result, LspProvider/Service, four operations, LspError
skills.md	the skill service: discovery priority, SkillSummary/SkillDefinition, session-prefix catalog, model-facing skill loading
compaction.md	the compaction seam: the compact/* session events, CompactionResult, the CompactService interface
subagent.md	the subagent seam: the named-provider registry, SubagentStartRequest/Result/Run, the start-time-vs-runtime capability split
web.md	the web access seam: WebSearchRequest/Result, WebFetchRequest/Result, WebFetchBody, provider availability, WebError
spill.md	the spill storage seam: SaveTextSpill, SpillOwner/SpillSource, SpillRef, the branded SpillLocator
workflow.md	the workflow seam: WorkflowStartRequest, WorkflowMeta, WorkflowRun/Result, the workflow/* event payloads, WorkflowError fatality

Type declarations and their JSDoc on these pages are source-equivalent and drift-checked by pnpm run verify-type-equiv (see development.md). Ordinary blocks preserve complete declarations; public-api blocks preserve body-stripped public class declarations. Cordis services use the generated service catalog.

The …Map → derived-union pattern

Almost every extensible sum type in the harness follows one shape: an interface keyed by a discriminant tag (the …Map), from which the union is derived with keyof. Plugins add variants by declaration merging — no edit to the owning package.

// The pattern, schematically:
interface ThingMap {
  'a': { kind: 'a'; /* … */ }
  'b': { kind: 'b'; /* … */ }
}
type ThingKind = keyof ThingMap          // 'a' | 'b'
type Thing = ThingMap[keyof ThingMap]    // the discriminated union

// A plugin extends it without touching the source package:
declare module '@deepseek-ai/dsh-llm' {
  interface ThingMap {
    'c': { kind: 'c'; /* … */ }
  }
}

Five canonical maps use this pattern; a plugin author extends these:

Map	Package	Derives	Catalog
ContentBlockMap	dsh-llm	ContentBlock	below
MessageSourceMap	dsh-llm	MessageSource	below
FinishReasonMap	dsh-llm	FinishReason	below
TurnEndReasonMap	dsh-session	TurnEndReason	session.md
SessionEventMap	dsh-session	SessionEvent	session.md

Two large discriminated unions are the ones consumers switch over most: StreamChunk (the streaming protocol) and SessionEvent (the log entry). Per the repo convention, switch on the tag — don't chain ifs — so each arm narrows and a typo'd tag fails to compile.

Branded IDs

IDs that cross package boundaries are branded — structurally strings, but non-interchangeable at the type level (a SessionId cannot be passed where a CallId is expected). Construction goes through a per-type factory; comparison, logging, and JSON behave as ordinary strings.

The Branded<B> primitive lives in its own type-only package, dsh-brand (no runtime code, no harness-package dependency), so any package can brand the ids it owns without depending on an unrelated capability package.

Source: packages/util/brand/src/index.ts

/** A string carrying a compile-time-only brand `B`. */
type Branded<B extends string> = string & { readonly [BRAND]: B }

The two core IDs are CallId (correlates a tool call with its result; dsh-llm) and SessionId (the shared live agent and durable session identity; dsh-session). Capability packages brand their own ids too, such as TaskId in tasks.md.

Content blocks and messages

A conversation is Messages; a message is an array of typed content blocks. The block union derives from ContentBlockMap.

Source: packages/llm/llm/src/types.ts

/**
 * Merge-extensible content blocks keyed by `type`. New core blocks must land
 * with adapter, UI, and compaction support.
 */
interface ContentBlockMap {
  'text': TextBlock
  'reasoning': ReasoningBlock
  'tool-call': ToolCallBlock
  'tool-result': ToolResultBlock
}

The block interfaces (full fields in source): TextBlock (text), ReasoningBlock (thinking, distinct from visible text), ToolCallBlock (id: CallId, name, raw-JSON arguments), ToolResultBlock (toolCallId, nested content: ContentBlock[], isError?). ContentBlock = ContentBlockMap[ContentBlockType]. The core set is limited to blocks every shipping path honors — multimodal content (images, audio, …) has no core block type; a feature that needs one adds it via the merge-extensible map together with the adapter/UI/compaction support that honors it.

Source: packages/llm/llm/src/message.ts

A Message is one identified, immutable role/source/content value. Model-produced assistant messages carry provider/model ownership and optional adapter-private replay metadata in their source:

/** Provider ownership and adapter-private replay data for an assistant message. */
interface AssistantProvenance {
  /** Provider route that produced the message. */
  provider: string
  /** Provider model id that produced the message. */
  model: string
  /**
   * Lossless-JSON adapter state needed to replay the provider response.
   * `LlmService` exposes it to a target adapter only when that adapter instance
   * currently owns both this historical provider and the target provider.
   */
  replayState?: unknown
}

/** One immutable message representation shared by delivery, durable history, and model requests. */
interface Message {
  /** Stable identity preserved across every representation boundary. */
  readonly id: MessageId
  /** Provider-neutral conversation role. */
  readonly role: 'system' | 'user' | 'assistant'
  /** Exact model-facing blocks. */
  readonly content: ContentBlock[]
  /** Required producer provenance. */
  readonly source: MessageSource
}

Where a message came from is itself a merge-extensible sum type:

/**
 * Where a message (or injected content) came from.
 * Merge-extensible sum type — plugins add their own `kind`s.
 */
interface MessageSourceMap {
  user: { kind: 'user' }
  plugin: { kind: 'plugin'; plugin: string }
  model: ModelMessageSource
  tool: ToolMessageSource
}

Streaming

Adapters emit a raw chunk protocol; the loop logs the chunks (replay fidelity) while feeding the same chunks through a BlockAssembler to rebuild blocks and messages. StreamChunk is a closed discriminated union over type — block-start, text-delta, reasoning-delta, tool-call-delta, block-end, usage, finish.

The full union, the adapter contract (usage-before-finish, raw-JSON tool arguments, the two sanctioned error paths), and BlockAssembler live on llm-streaming.md.

The model request

One model call is a fully-assembled GenerateOptions. The adapter answers with a raw StreamChunk stream; the consumer assembles it with BlockAssembler (see llm-streaming.md).

Source: packages/llm/llm/src/types.ts

Provider and model discovery uses small provider-neutral descriptors. A model catalog is advisory: routing still keys on a registered provider, and an adapter may accept unlisted model ids.

Registering an adapter returns a handle: the disposer, plus the atomic route replacement a plugin whose route set is user-configurable needs.

/**
 * What {@link LlmService.registerAdapter} returns: the disposer, plus an
 * atomic route replacement for the same adapter instance.
 */
interface AdapterRegistrationHandle {
  /** Release every route this registration currently holds. */
  (): void
  /**
   * Replace this registration's routes with `providers`, keeping the same
   * adapter instance. The candidate set is validated in full first — a
   * conflict with another adapter, an invalid name, or bad provider metadata
   * throws and leaves the current routes untouched — and the swap itself is
   * one synchronous section, so no request can observe a gap. An empty array
   * is legal here (a settings section that emptied holds zero routes while
   * staying registered), unlike an empty initial registration.
   *
   * Throws `LlmError` with code `REGISTRATION_DISPOSED` once the registration
   * has been released: its routes are gone and its disposer has already run,
   * so anything registered afterwards would have no owner left to release it.
   * @param providers - the complete next route set for this registration.
   */
  replace(providers: string[]): void
}

/** Display metadata for one registered provider route. */
interface LlmProviderInfo {
  /** Provider route key used by {@link GenerateOptions.provider}. */
  id: string
  /** Human-readable provider name for selectors and diagnostics. */
  name: string
}

Adapter plugins additionally declare which routes could run through registerConfigurableProviders(), addressing each one's user-settings section, so configuration surfaces can offer dormant providers before any route registers.

/**
 * One provider route an adapter plugin can activate through configuration,
 * whether or not the route is currently registered. Configuration surfaces
 * merge this directory with `listProviders()` to offer every configurable
 * provider alongside its live/dormant state.
 */
interface LlmConfigurableProvider {
  /** Provider route key this entry activates when configured. */
  provider: string
  /** Human-readable provider name for configuration surfaces. */
  displayName: string
  /** User-settings namespace whose section configures this provider. */
  settingsNs: string
  /**
   * Path from that namespace's section root to this provider's profile
   * object; empty when the whole section is the profile.
   */
  settingsPath: readonly string[]
}

/** One adapter-discovered model; catalog membership is advisory, not request validation. */
interface LlmModelInfo {
  /** Provider route that owns this model entry. */
  provider: string
  /** Model id passed to {@link GenerateOptions.model}. */
  id: string
  /** Human-readable model name for selectors. */
  name: string
  /** Optional user-facing distinction from otherwise similar models. */
  description?: string
}

Correctness-sensitive metadata is resolved separately from the advisory catalog and is owned by the adapter serving the exact route. Context capacity, adapter call defaults, and reasoning choices share one exact-model result so consumers do not repeat authoritative model resolution.

/** Provider-owned context capacity for one exact provider/model route. */
interface LlmModelContext {
  /** Maximum combined request and response context in tokens. */
  contextWindow: number
}

Reasoning effort is another exact-route capability. The core brands identifiers but does not enumerate their values; each adapter owns the ordered set, display names, and optional deployment default.

/** Adapter-owned identifier for one model's selectable reasoning effort. */
type ReasoningEffortId = Branded<'ReasoningEffortId'>

/** Display metadata for one adapter-owned reasoning effort. */
interface LlmReasoningEffortInfo {
  /** Opaque stable value accepted by {@link GenerateOptions.reasoningEffort}. */
  id: ReasoningEffortId
  /** Human-readable effort name for selectors and diagnostics. */
  name: string
  /** Optional user-facing distinction from otherwise similar efforts. */
  description?: string
}

/** Selectable reasoning efforts for one exact provider/model route. */
interface LlmModelReasoningInfo {
  /** Supported efforts in adapter-preferred display order. */
  efforts: readonly LlmReasoningEffortInfo[]
  /**
   * Adapter-configured default materialized into requests when callers omit
   * an effort. Absence preserves the provider's own default.
   */
  defaultEffort?: ReasoningEffortId
}

/** Exact-route model metadata resolved by its owning adapter. */
interface LlmResolvedModelInfo extends LlmModelInfo {
  /** Provider-owned context capacity when known. */
  context?: LlmModelContext
  /** Adapter-configured per-request output cap materialized when callers omit one. */
  defaultMaxTokens?: number
  /** Adapter-owned selectable reasoning levels when exposed. */
  reasoning?: LlmModelReasoningInfo
}

/** A single model request, fully assembled. */
interface GenerateOptions {
  /** Registered provider route selecting the adapter instance. */
  provider: string
  model: string
  /** Adapter-owned reasoning effort selected for this exact model. */
  reasoningEffort?: ReasoningEffortId
  /**
   * Ordered conversation messages, exactly as the provider sees them (after
   * the `system` slot). A loop-built request assembles them as
   * the derived history (dsh-agent-loop); a hand-built one-shot passes any list.
   */
  messages: Message[]
  /** System prompt text (adapters map to the provider's system slot). */
  system?: string
  /** Tool schemas (adapters map to the provider's `tools` field). */
  tools?: ToolSchema[]
  temperature?: number
  maxTokens?: number
  /**
   * Stop sequences: generation halts as soon as the model produces any one of
   * these strings (adapters map to the provider's stop field, e.g. OpenAI
   * `stop`). The stop string itself is not included in the output.
   */
  stop?: string[]
  signal?: AbortSignal
  /**
   * Session identity stamped by the loop for listener routing. Adapters ignore
   * it; replay uses it to keep concurrent parent and child cursors independent.
   */
  sessionId?: Branded<'SessionId'>
  /**
   * Provider-neutral classification for an auxiliary model call. Adapters may
   * map the purpose to model-hidden transport metadata or purpose-specific
   * generation policy. Ordinary conversation requests leave it unset.
   */
  purpose?: 'compaction' | 'session-title'
}

Why a model response stopped is a merge-extensible reason. Terminal provider failures carry the streaming contract's LlmFailure:

/**
 * Why a model response stopped.
 * Merge-extensible so adapters can surface provider-specific reasons.
 */
interface FinishReasonMap {
  'stop': { kind: 'stop' }
  'tool-calls': { kind: 'tool-calls' }
  'max-tokens': { kind: 'max-tokens' }
  'aborted': { kind: 'aborted'; failure: LlmFailure }
  'error': { kind: 'error'; failure: LlmFailure }
}

FinishReason = FinishReasonMap[keyof FinishReasonMap]. TokenUsage (per-call accounting with disjoint cache fields) is detailed on llm-streaming.md.

GenerateOptions.tools carries ToolSchema — the JSON-schema description of a tool, as sent to the model. It is declared in dsh-llm (not dsh-tools) precisely because it is part of the request the loop assembles every step:

/**
 * JSON-schema description of a tool, as sent to the model.
 *
 * Declared here (not in dsh-tools) because it is part of {@link GenerateOptions};
 * dsh-tools' ToolDefinition and dsh-system-prompt's PromptAssembly both import
 * it from this package.
 */
interface ToolSchema {
  name: string
  description: string
  /** JSON Schema object for the arguments. */
  parameters: Record<string, unknown>
}

The model-facing ToolSchema is the wire shape; the registered ToolDefinition that produces it (schema + execute) is on tools.md.

The request envelope: LlmCallConfig and the logged header

The loop builds each request from logged state. EpochHeader records call config, adapter-default provenance, rendered prompt, and authoritative returned tool order (configured by toolOrder, or lexicographic when unset) through full request/header snapshots. Together with derived history, this makes the request reconstructable from the session log. See session.md and the reconstructability Agent Note.

agent/request receives a frozen call-config seed and may return a replacement to switch provider, model, reasoning effort, or sampling. Before the waterfall, the loop removes values marked as adapter defaults so exact-model preparation materializes the selected route's current values; unmarked explicit settings remain in the proposal. After the waterfall, preparation rejects unsupported explicit effort ids without clamping and logs the effective config plus provenance under the turn signal. The prepared call keeps one adapter registration through dispatch. Requests reaching llm/stream are deep-frozen, so mutation throws, and carry a process-local loop identity so observers do not confuse separately logged frozen auxiliary calls with conversation requests.

On the wire, a loop-built request reads the system slot (the rendered prompt assembly) followed by the derived history — the boundary snapshot, whose tail is the newest user/message on a turn's first step and the previous step's tool results on later steps. The dev invariant recomputes exactly this equation against every loop-built request.

FIXME(call-config-shape): revisit which remaining fields are genuinely epoch-level for cache purposes (model and the model-owned reasoning effort are explicit; the sampling scalars sit here out of caution).

/**
 * Provider, model, reasoning effort, and sampling scalars of one conversation's
 * requests. Every field maps 1:1 onto the same-named `GenerateOptions` field;
 * the loop builds requests from the logged header rather than accepting these
 * per call.
 */
interface LlmCallConfig {
  provider: string
  model: string
  reasoningEffort?: ReasoningEffortId
  temperature?: number
  maxTokens?: number
  stop?: string[]
}

/**
 * Effective config fields supplied by exact-model adapter resolution rather
 * than by the caller's request proposal.
 */
interface LlmCallConfigAdapterDefaults {
  reasoningEffort?: true
  maxTokens?: true
}

Sessions

A Session is an append-only log of typed SessionEvents — the single source of truth. The LLM message history is derived from the log (deriveMessages()), not stored separately. The event vocabulary derives from SessionEventMap:

Source: packages/core/session/src/types.ts

/**
 * One immutable entry in the session log.
 *
 * A proper discriminated union over `type` (not independent `type`/`data`
 * unions), so `switch (event.type)` narrows `event.data` without casts.
 *
 * The {@link sourceEventSeqs} and {@link surfaceOp} fields are conditional:
 * they only exist on {@link SurfaceEventType} variants (`user/message`,
 * `assistant/message`, `tool/result`, `steering/message`).
 * Non-surface events (boundary markers, chunks, usage, errors) never carry
 * surface metadata — the compiler enforces this at `Session.append()`
 * call sites.
 */
type SessionEvent<T extends SessionEventType = SessionEventType> = {
  [K in SessionEventType]: {
    type: K
    /** Monotonic sequence number within the session. */
    seq: number
    /** Unix epoch milliseconds. */
    time: number
    data: SessionEventMap[K]
  } & (K extends SurfaceEventType ? {
    /**
     * Seq numbers of events that are provenance sources of this event
     * (e.g. the `assistant/chunk` seqs that built an `assistant/message`,
     * or the surface nodes shadowed by a compaction replace node). An
     * `assistant/message` may carry a present empty array for a known empty
     * provider stream; omission means unrecorded provenance.
     */
    sourceEventSeqs?: number[]
    /** How this event entered the surface; absent for non-surface events. */
    surfaceOp?: SurfaceOp
  } : object)
}[T]

The session event variants, deriveMessages() projection rules, TurnEndReason vocabulary, and execution-enclosure and standalone-event rules are on session.md. How the log is made durable — the SessionPersistence seam, JSONL/SQLite backends, the session/flush checkpoint, crash recovery, and SessionHeader — is on persistence.md.

The agent handle

Agent is the surface every plugin (UI, hooks, orchestrators) programs against. The concrete implementation is package-internal to dsh-agent-loop; nothing outside the loop depends on it.

Source: packages/core/agent/src/types.ts

/** One of the two ordered pending-message lists owned by an agent. */
type InboxTarget = 'next-turn' | 'next-step'

Every pending occurrence is its UserMessage; MessageId is the sole identity. Inbox.append, prepend, update, remove, clear, and splice record normalized durable agent/inbox/spliced mutations and reject duplicate pending ids. Ordinary removals and clear() are cancellations. claim(target) atomically removes the proposed step batch through pure deletion splices; the loop separately emits per-message claimed notifications. Whole-queue consumers such as UI projections reconstruct nextTurn and nextStep from the durable splices, while consumers following one message use the exact agent/inbox/inserted, claimed, and discarded notifications.

/** Options for {@link Agent.cancel}. */
interface CancelOptions {
  /**
   * Preserve queued and steering inbox items instead of discarding them. The
   * active turn is still aborted, but un-started and pending work survives for a
   * later turn and no canceled inbox splice is logged.
   */
  keepInbox?: boolean | undefined
}

/** Why an active agent driver was cancelled. */
type AgentCancelCause =
  | { readonly kind: 'user' }
  | { readonly kind: 'parent' }
  | { readonly kind: 'hook'; readonly reason: string }
  | { readonly kind: 'disposed' }

Agent is an interface over the public live-agent contract. Its unified send method exposes target and wakeup routing directly; followup, steer, and inject are fixed-preset aliases.

/** Public live-agent handle. */
interface Agent {
  /** The single identity shared with {@link session}. */
  readonly id: SessionId
  /** The provider route and model this agent's requests use. */
  readonly options: AgentOptions
  /** The live session this agent drives; its log is the durable source of truth. */
  readonly session: Session
  /** The agent-owned projection of durable pending work. */
  readonly inbox: Inbox
  /** The current lifecycle state, mirrored on every `agent/status` transition. */
  readonly status: AgentStatus
  /** Agent-scoped context; its contributions are agent-local, unwind on disposal, and reject registration afterward. */
  readonly ctx: Context

  /**
   * Clear queued and steering work — unless `keepInbox` — and abort the active
   * turn. The first cause wins for the active turn. Idle cancellation is a
   * no-op and does not arm later work.
   * @param cause - the stable caller intent carried by the current turn signal.
   * @param options - cancellation options; `keepInbox` preserves pending work.
   */
  cancel(cause: AgentCancelCause, options?: CancelOptions): void

  /**
   * Resolve after the current whole-agent activity reaches quiescence. This
   * follows replacement work scheduled before the observed driver retires,
   * but does not identify the settlement of any particular message.
   * @returns fulfillment after no scheduled or active driver remains.
   */
  whenIdle(): Promise<void>

  /**
   * Route identified input to an inbox boundary and optionally wake the driver.
   * Waking input submitted after active cancellation is queued for the next turn.
   * @param message - identified content and its producer provenance.
   * @param target - the preferred next-turn or next-step inbox boundary.
   * @param wakeup - whether delivery may wake the driver.
   */
  send(message: UserMessage, target: InboxTarget, wakeup: boolean): void

  /**
   * Queue an ordinary follow-up turn and wake the driver. The item becomes the
   * sole ordinary message of its own turn.
   * @param message - identified prompt content and its producer provenance.
   */
  followup(message: UserMessage): void

  /**
   * Submit steering for the nearest step. An idle driver schedules a turn;
   * collecting and running drivers consume it at their next step boundary.
   * Cancellation or disposal may discard pending steering.
   * @param message - identified steering content and its producer provenance.
   */
  steer(message: UserMessage): void

  /**
   * Queue model-facing context for the next pre-step without waking the
   * driver. Collecting and running drivers claim it at the nearest later
   * step boundary; idle drivers leave it pending until follow-up or steering
   * wakes them. It may miss a request whose pre-step already claimed its
   * batch. Cancellation or disposal may discard pending context.
   * @param message - identified injected context and its producer provenance.
   */
  inject(message: UserMessage): void
}

AgentStatus is 'idle' | 'running', and SessionId is branded. Disposal removes the agent from the registry and emits agent/disposed; it is not a terminal status value. running describes the driver-wide drain interval and may span consecutive queued turns; it does not prove a turn is still open. followup() returns no handle: its MessageId identifies durable inbox insertion, claim, and discard facts, not a later assistant output or turn ending. whenIdle() observes the whole agent, so callers may call a receipt-to-idle interval a run only when they explicitly own that interval (decision). AgentOptions is merge-extensible: core declares provider?, model?, and maxTokens? (dispatch requires provider and model after agent/request). When present, maxTokens must be a positive safe integer and caps every conversation-model request; omission allows the exact-model adapter default to materialize before the request header, or otherwise leaves provider behavior unchanged. Persona belongs to dsh-system-prompt: an agent-scoped deployment:persona may shadow the global default.

The cause is a TypeScript-enforced same-process input. An active cancellation holder copies it into the runtime-only AbortSignal.reason; a signal grants cooperating listeners no classification authority. Durable turn/end retains the coarse { kind: 'aborted' } outcome; request provenance would require a separate durable event rather than overloading the terminal result.

The event taxonomy owns the agent/* lifecycle, checkpoint, and waterfall contracts. Turn and step boundaries are durable session events rather than agent emits.

Initiating Agent

The process-local initiator carried by ctx.agents is the exact Agent above, not a separate frame or copied identity. Ambient presence is neither liveness proof nor authorization; the initiator-scope decision owns its lifetime and boundary rules.

Interception decisions

Pre-step decisions use the same identified UserMessage shape as durable user-role input. The entered batch is authoritative and preserves every message's identity and provenance. Hook bridges map their native decision fields onto this typed result.

Source: packages/core/agent/src/types.ts

agent/pre-step receives the exclusive claimed batch and the proposed step's coordinates and cancellation signal. The initial proposal runs before its turn opens; a tool continuation may submit an empty claimed batch between steps:

/** Coordinates and cancellation for a proposed step. */
interface PreStepContext {
  /** Turn that will own the step. */
  readonly turn: number
  /** Step proposed by the loop. */
  readonly step: number
  /** Current turn cancellation signal. */
  readonly signal: AbortSignal
}

It returns a PreStepDecision. Reject opens no step. Enter supplies the complete message batch appended after step/start; claimed messages omitted by the final decision remain removed, while input inserted after the claim stays pending:

/** Whether and with which messages the loop enters a proposed step. */
type PreStepDecision =
  | { kind: 'reject' }
  | { kind: 'enter'; messages: UserMessage[] }

agent/request-error runs after a failed model step closes and before its turn closes. Listeners can repair durable state or await policy work while the failed turn's signal is still live. A handling listener returns { kind: 'retry' } without calling next(); the default undefined leaves the failure terminal.

/** Action returned by a listener that owns model-request recovery. */
type RequestErrorAction = { kind: 'retry' } | undefined

agent/pre-step is the single serial boundary before request derivation. agent/turn-stopping runs when a turn has no tool or steering continuation, before one final steering drain.

agent/session-start carries a SessionStartSource (why the session lifecycle began; a bridge keys its SessionStart matcher on it):

/** Why a session lifecycle began; seeded creates are `startup`, while persisted loads are `resume`. */
type SessionStartSource = 'startup' | 'resume' | 'clear' | 'compact'

ToolDefinition

The one pipeline-authoring type that is core: what every registered tool is — a model-facing ToolSchema plus an execute function and optional final-content and UI callbacks. A tool author rarely constructs it by hand (the defineTool DSL builds it with typed args), but it is the contract the registry holds and the loop dispatches through.

Its full fields, the defineTool/ValueSchemaSpec/ParameterSchemaSpec typed schema DSL, the ToolExecution/ToolExecutionResult waterfall shapes, and the tool-presentation UI vocabulary are on tools.md.