deepseek-harness/packages/core/agent-loop/README.md

---

description: "The default agent driver for users and maintainers choosing, configuring, or debugging how agents are created and how turns and steps run."

kind: "package-reference"

@deepseek-ai/dsh-agent-loop

English | 中文

Summary

dsh-agent-loop creates fresh agents or resumes persisted sessions, then drives each turn through model requests, streamed responses, tool execution, and durable session history. Mount it for standard agent compositions; declarative entries start agents at boot, while the public ctx.agents API supports programmatic creation and resume. maxParallelToolCalls limits concurrent parallel-safe calls, and exclusive calls retain ordering. Cancellation preserves streamed text already delivered to the user. Choose a custom Agent implementation only when the standard "call model, run tools, repeat" lifecycle is insufficient.

Table of Contents

• Use this package
• Understand the implementation
• Further Exploration
• Model Experience
• Known Limitations and Deferred Work
• Dev Note

---

Use this package

Mount dsh-agent-loop in any composition that should run agents. It supplies the driver behind ctx.agents and starts any agents you declare in its config; both dsh-base and dsh-sdk-minimal mount it as an explicit row.

Configure declarative agents

Agents declared in the config start automatically when the plugin loads. Each entry needs an id label; a model call additionally requires both provider and model (agent/request may supply a missing pair before dispatch).

- name: '@deepseek-ai/dsh-agent-loop'
  config:
    maxParallelToolCalls: 10
    agents:
      - id: 'main'
        provider: deepseek
        model: deepseek-chat
        reasoningEffort: high
        cwd: /workspace

Field	Default	Meaning
maxParallelToolCalls	10	Parallel-safe tool calls in flight per step; 1 is serial
agents[].id	required	Stable label; a fresh session mints ${id}-session-<uuid> unless sessionId is set
agents[].provider / agents[].model	—	Model route; both required before dispatch
agents[].reasoningEffort	—	Non-empty initial reasoning effort; agent/request may override it
agents[].maxTokens	—	Positive per-request output-token cap
agents[].cwd	—	Workspace directory for a fresh session
agents[].sessionId	—	Exact identity: first use creates, a remount resumes materialized history
agents[].resumeSessionId	—	Load this persisted session instead of creating one; mutually exclusive with sessionId

The generated configuration catalog is the exhaustive source for every accepted field. The adapter validates the effective reasoning effort and the loop records it in the request header. maxParallelToolCalls is also the whole agent-loop settings section, so a user layer over this entry caps the next tool group without a restart.

Create or resume agents programmatically

Plugins and hosts create agents through ctx.agents.create() and resume persisted sessions through ctx.agents.resume(); both return an AgentHandle whose dispose() owns exact teardown. The loop runs each created agent to completion — the handle is only needed when the caller must tear the agent down itself.

const handle = await ctx.agents.create({
  sessionId,
  agentOptions: { provider: 'deepseek', model: 'deepseek-chat' },
  setup: (agentCtx, agent) => { /* scoped registrations plus explicit unpublished Agent */ },
})

Every inbox mutation commits one normalized agent/inbox/spliced event. The projection registry folds that event synchronously, so the live projection reflects the splice when Session.append() returns. Insertions, edits, removals, claiming, and cancellation replay through the same standard splice coordinates. Ordinary removals carry outcome: 'canceled' and emit agent/inbox/discarded { message }; claiming uses pure deletions with no outcome and emits agent/inbox/claimed. Every insertion emits agent/inbox/inserted { message }. MessageId stays unique across both pending lists. Consumers that need a removed message use the claimed or discarded notification instead of depending on a pre-splice session/event view.

What a step does

Each step sends the session's derived history — with the latest non-empty system/message node as the effective prompt, or no system messages when the rendered prompt is empty — and its visible tool schemas; the model's tool calls run through the guarded tool pipeline and every accepted fact is appended to the session log before the next step derives from it. Parallel-safe calls may overlap up to maxParallelToolCalls; exclusive calls run alone as ordering barriers. Cancellation is cooperative: agent.cancel() aborts the current activity and, unless keepInbox is set, clears pending work; a cancelled stream finalizes the text already delivered to the user.

---

Understand the implementation

Implementation internals — click to expand

This section explains how the package realizes the behavior above; the observable contract is covered in [Use this package](#use-this-package). ### Design concept The package is the one concrete implementation of the public `Agent` contract. It registers itself as the `AgentFactory` on `ctx.agents`, so consumers never import this package; ownership of each created agent lives with the caller fiber and the loop provider, converging on one memoized quiescence boundary. Every observable effect happens through session events and the `agent/*` taxonomy — package internals are never part of the public surface. ### Request headers and adapter defaults After `agent/request`, `ctx.llm.prepareCall()` validates adapter-owned fields and resolves reasoning-effort and output-token defaults under the active turn signal. The loop retains that exact adapter through resolution, `request/header` logging, and dispatch. It writes a full header for the first request, a changed envelope (config or tools — the prompt is not part of the header), an explicit message-series start, a request after surface replacement (an in-place prompt replacement or compaction), and resume; unchanged steps, retries, and ordinary later turns in the same series inherit the latest header, and an in-history prompt append is not a replacement, so the request that follows it inherits the header too. Beside the header, the loop logs `request/context` — provider, model, `contextWindow`, and the route's `systemPromptUpdate` mode from `prepareCall()` — only when one of those differs from the latest snapshot. Before the next waterfall, the loop removes adapter-default fields so the current route resolves them again, while explicit settings persist. An unhandled route still fails with `NO_ADAPTER`. The loop deep-freezes each derived message identity on its first request and reuses that proof only within the same agent. Restored messages keep their identity; request construction does not freeze their containing event wrappers. Each request freezes its local canonical header, fresh message array, and envelope while leaving the cancellation signal live. The [request-freeze decision](../../../.agents/notes/implemented/simplification/2026-09-06-agent-request-freeze-provenance.md) explains ownership and measurement. ### Source map | File | Role | |---|---| | [`src/index.ts`](src/index.ts) | Plugin entry: `AgentLoop` service, config schema, declarative agent startup, factory registration | | [`src/agent.ts`](src/agent.ts) | The concrete `ReactLoopAgent` driver: inbox, turn/step machine, cancellation | | [`src/inbox.ts`](src/inbox.ts) | Package-internal `ReactLoopInbox`: durable projection, structural commands, and loop-only claim state | | [`src/tool-calls.ts`](src/tool-calls.ts) | Tool scheduling: exclusive barriers and the bounded parallel pool | | [`src/runtime-context.ts`](src/runtime-context.ts) | Per-step runtime-context snapshot handling | | [`src/constants.ts`](src/constants.ts) | `DEFAULT_MAX_PARALLEL_TOOL_CALLS` | | [`src/invariant.ts`](src/invariant.ts) | Invariant companion: request reconstruction from the session log | ### Creation and teardown Creation is one rollback-covered transaction: construct a private session, concrete agent, and scoped context; await optional setup with the context and Agent passed separately; enter both registries; announce `session/created` then `agent/created`; emit `agent/session-start`; only then start the driver. A caller creating a runtime child sets `options.parentAgent`; the caller Context separately owns the transaction and live handle. A setup throw, commit failure, or owner disposal rolls the transaction back without publishing either id. Teardown runs stop-and-drain, closes the session's write path, unwinds the scope, detaches the agent, then detaches the session, and every detach is bound to the exact entered object so a stale disposer cannot remove a later same-id replacement. ### Persistence integration The loop is the production acquisition point for session write handles. When `ctx.sessionPersistence` is mounted, `create`/`createAgent` call `persistence.create(header)` — storing the durable identity and taking write ownership before publication — and append the constructor seed through the handle; `resume` calls `persistence.open(id, 'write')` first (excluding a concurrent resume of the same id), reads the physically valid log through the handle, and appends `interruptedTurnClosers` for a log crashed mid-turn as an ordinary batch — semantic crash repair is the agent layer's job, not a storage entry point. Immediately before publication, `appendUnstoredSuffix` stores any events appended during the setup window (seed markers, delegation policy records), which never re-emit through `session/event`. Once published, the mounted backend routes the session's `session/event` batches, `session/flush` barriers, and `session/disposed` retirement into the active write handle by session id; the loop touches storage only through the handle it owns. The memoized teardown closes the handle — close drains any routed buffer — after the loop commits the session's closing events, provably releasing write ownership. Without a backend, sessions are memory-only and nothing else changes. ### Turn and step flow The driver owns one agent for its lifetime and runs inside `ctx.agents.withInitiator(agent, ...)`. Its package-internal `ReactLoopInbox` constructor registers the standard `inbox` projection on the agent scope, then uses that projection for structural commands and loop-only claims. Registry reference counting keeps the shared key active until the last agent scope unloads. At a turn boundary it opens the durable turn, then atomically claims pending next-step input plus one queued prompt; between steps it claims only next-step input. The driver assembles the prompt and tools, projects runtime context, and runs `agent/pre-step`. A rejected decision or empty first batch opens no step. On the first attempt after acceptance, `step/start` precedes the `agent/request` waterfall and `prepareCall()`; neither async phase sees the pending system prompt or accepted users committed to history, and cancellation during either commits neither. For every attempt, the loop then synchronously reconciles the rendered prompt against the surviving `system/message` nodes using the prepared call capability, appends the accepted `user/message` batch only on the first attempt, logs the header and context as needed, and derives and freezes the request before streaming through that bound prepared call. Retries reuse the same rendered assembly without repeating assembly, `agent/pre-step`, or user admission. Reconciliation sees pre-step and retry compaction; a broken series consolidates the prompt at the head rather than appending an update after already-committed users. The request is `header.config`, `deriveMessages()`, and `header.tools`; it carries no `system` field. Each model attempt emits one process-local `start`, emits every `chunk` only after the matching durable assistant-frame settlement, and emits exactly one terminal `end`; final assembly or message-append failure settles it as `aborted`, while `committed` follows the durable `assistant/message`. Each successful model call appends one message anchor, and a cancelled stream appends an `interrupted: true` anchor with the delivered prefix so the next request contains what the user saw. Within a step, exclusive calls form barriers and parallel-safe calls use the bounded rolling pool; policy, durable results... (line truncated to 2000 chars) Prompt admission uses the actual `prepareCall()` result, not the preceding `request/context`. With no system node, even an empty prompt is appended (reserving node 0 without a wire message). On an incapable route or at a new request series, a non-empty rendering is consolidated at the first system node: each non-empty later system node receives a logged per-node empty replacement, then the head is rewritten if needed. Dormant empty tails need no replacement and do not determine the effective text. Consolidation applies even when the latest effective text is unchanged. On a continuing `in-history` series, an unchanged effective prompt produces no event and a non-empty change appends. An empty rendering clears every non-empty later system node through a logged per-node empty replacement, then empties the head if needed, regardless of route or series state. No older instructions remain model-visible. An empty head with no active later system node means no prompt; repeated clears and resume leave it empty. A restored non-empty prompt follows the same route/series rule: a continuing capable route may append it, while an incapable route or new series refills the head. A step starts a series when the pre-step decision declares `startsRequestSeries`, when the surface replace generation changed since attachment or the last request (compaction or any replacement), or when visible tool schemas changed. Resume and a provider or model swap alone continue the series; the prepared route still governs admission. Per-node empty replacements preserve intervening history without a surface delete operation. ### Failure and cancellation Final adapter selection, dispatch, and iteration failures arrive as terminal finishes and enter `agent/request-error`; a handling listener returns `{ kind: 'retry' }` without calling `next()`, while an unhandled failure is terminal. Middleware, result-processing, tool, and other extension failures remain thrown and close the turn directly — plugin failure ends the turn, not the loop. Undispatched model tool calls after cancellation receive synthetic `tool/call` plus `ABORTED_BEFORE_DISPATCH` result pairs. The [explicit-cancellation decision](../../../.agents/notes/implemented/architecture/2026-07-16-explicit-turn-cancellation.md) owns the signal lifecycle.

---

Further Exploration

The package-level contract is enough for most consumers; read these when you need the surrounding domain and the design rationale.

• agent package — the Agent handle, registry, and agent/* events this loop implements.
• Core subsystem — the turn flow and interception decisions.
• Session subsystem — the durable log the loop writes and derives from.
• Tools subsystem — the pipeline the loop dispatches through.
• Explicit-cancellation Agent Note — signal lifetime and cancellation races.
• Core group map — how the core packages compose.

---

Model Experience

Complete conversation request

What the model sees

For each step, the loop sends the session's derived messages and visible tool schemas. Non-empty system/message nodes carry the prompt, with the latest as the effective version; an empty rendering clears all prompt versions from derived history. It supplies provider, model, and cwd variable values but no additional fixed prose.

Token effect

System text and schemas are paid again on every step, and on an in-history route every retained prompt version is paid until compaction shadows it or prompt reconciliation empties it. Per-agent scoping chooses the contributions, while the authoritative assembly waterfall can alter the final request and makes its listener responsible for protocol coherence.

KV Cache effect

Append-only only while system text, schemas, and earlier history remain byte-identical under the same provider and model route. An unchanged rendered prompt keeps the cached prefix unless an incapable route or a new request series must consolidate retained in-history system nodes. A prompt change that replaces a system node in place makes the request differ from that node's first token — in full when the node is node 0 — so the provider prefix cache misses from there; when the prepared call declares systemPromptUpdate: 'in-history', a non-empty prompt change inside a continuing request series is appended after the cached history, so the prefix through that history stays reusable. A schema or composition change invalidates reuse from the first altered request token.

Retained message history

What the model sees

Accepted user messages, assistant messages, tool calls and results, injected context, and steering are logged and sent on later steps. Raw stream chunks, lifecycle boundaries, and other log-only events are excluded.

Token effect

Input grows with every surface message until a compaction replacement shadows older nodes; a multi-step tool turn resends the accumulated history each step.

KV Cache effect

Ordinary history growth is append-only and preserves reusable entries. A surface replacement or compaction invalidates reuse from the first shadowed history token.

Undispatched calls after cancellation

What the model sees

If a later request replays an aborted step, each tool call that cancellation prevented from dispatching has error code ABORTED_BEFORE_DISPATCH and result text Error: tool call aborted before dispatch.

Token effect

One fixed error result per skipped call remains in history until compaction shadows it.

KV Cache effect

Append-only; each synthetic result follows the reusable request prefix and does not invalidate existing KV-cache entries.

Known Limitations and Deferred Work

These limits define when the loop needs special care. They are current package constraints, not a task backlog.

• Classification is unary — calls whose safety depends on comparing siblings or resources must remain exclusive (rationale).
• Config labels are fresh by default — omitting sessionId creates a fresh ${id}-session-<uuid> on every startup; exact resume-or-create behavior requires an explicit stable sessionId, while resumeSessionId requires existing persisted history.
• Config agents have no per-agent persona field or setup hook — they use the deployment persona; scoped persona and tool composition are available only through the programmatic ctx.agents.create() / resume() factory options.
• No built-in turn budget — tool calls or steering continue the current turn; a policy that bounds runaway turns must cancel from an existing lifecycle extension point such as agent/turn-stopping.

Dev Note

Working context for maintainers — click to expand

None.