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

Tools

The tool pipeline of dsh-tools. core.md introduces ToolDefinition as the one pipeline-authoring type promoted to the spine and ToolSchema as the model-facing wire shape. This page owns the full ToolDefinition, the typed schema DSL that builds it, the guarded execution shapes, and the UI-presentation vocabulary.

Source: packages/core/tools/src/index.ts · packages/core/tools/src/schema.ts · packages/core/tools/src/presentation.ts

ToolDefinition — a registered tool

A ToolSchema (the model-facing fields) plus the execute function and optional UI presenters. The registry holds these; the loop dispatches calls through them. The registry's schemas() builds the model-facing ToolSchema[] by an explicit allowlist — execute/presentCall/presentResult must never leak into a model request.

interface ToolDefinition extends ToolSchema {
  execute(args: unknown, exec: ToolExecution): Promise<ToolExecuteReturn>
  /**
   * Cooperative tool-call timeout budget in milliseconds. Omit for no deadline.
   * Enforced by `@deepseek-ai/dsh-timeout-policy` (a `tools/execute` wrapper); it
   * is NEVER sent to the model — `schemas()` whitelists only name/description/
   * parameters. Declaring it asserts this tool forwards `exec.signal` to a
   * cooperative implementation that can reach quiescence when the signal aborts.
   */
  timeoutMs?: number
  /**
   * Optional: how to present the PENDING state of one call in a UI, derived from
   * the call's `args` (parsed arguments, `unknown` — the tool validates/narrows
   * its own input). Returns a {@link ToolCallView} (a `card`-tagged render intent),
   * or `undefined` (or omit the method) to fall back to a generic presentation
   * (title = tool name, raw args as input). Pure and side-effect-free: a UI may
   * call it during live streaming AND a session-log replay, so it must depend
   * only on `args`.
   */
  presentCall?(args: unknown): ToolCallView | undefined
  /**
   * Optional: how to present the COMPLETED state, given the same `args` and the
   * `result` (`execute`'s content + whether it errored). Returns a
   * {@link ToolResultView}, or `undefined` (or omit the method) to keep the
   * pending title and render the raw result content. Pure and side-effect-free
   * for the same replay reason.
   */
  presentResult?(args: unknown, result: ToolResult): ToolResultView | undefined
}

execute receives args: unknown — a raw ToolDefinition validates its own input. First-party tools don't write that by hand; they use defineTool, which validates and narrows for them.

The typed schema DSL

Plugin authors write per-property specs with a boolean required: true, and a type-level helper maps the spec to the execute argument type — zero casts. The DSL is machinery that types ToolDefinition; it is intentionally a sub-page detail, not core.

Source: packages/core/tools/src/schema.ts

interface SchemaProp {
  type: SchemaType
  /** Per-property required flag (NOT the JSON Schema top-level required array). */
  required?: true
  /** Human-readable description, surfaced in the JSON Schema as well. */
  description?: string
  /** Enum of allowed values (strings only). */
  enum?: string[]
  /** Default value. */
  default?: unknown
  /** Nested properties for type: 'object'. */
  properties?: SchemaSpec
  /** Items schema for type: 'array'. */
  items?: SchemaProp
}

type SchemaSpec = Record<string, SchemaProp>

SchemaType is the primitive union 'string' | 'number' | 'boolean' | 'object' | 'array'. InferArgs<S> maps a SchemaSpec to the TS argument type — required: true props become required keys, everything else genuinely optional:

type InferArgs<S extends SchemaSpec> = Simplify<
  & { [K in RequiredKeys<S>]: InferPropValue<S[K]> }
  & { [K in Exclude<keyof S, RequiredKeys<S>>]?: InferPropValue<S[K]> }
>

defineTool({ name, description, parameters, execute, … }) ties it together: parameters is a SchemaSpec, execute(args, exec) gets args: InferArgs<typeof parameters>, and the helper converts the spec to JSON Schema (schemaSpecToJsonSchema) for the wire and validates model-generated args (validateArgs) before the typed body runs. A mismatch throws ToolArgsError (code: 'INVALID_ARGS'), which the registry turns into an isError result so the model can self-correct. Why a custom DSL and not schemastery: tool parameters need JSON Schema (the LLM wire format), not validation/transformation — the lightweight DSL gives the best authoring DX with the smallest surface.

Registration is a value boundary. ToolRegistry.register() reads every top-level field once, validates fixed scalar and callback types, materializes ToolDefinition.parameters as detached lossless JSON in one recursive pass, binds the accepted execute/presentation callbacks once to the original definition as their method receiver, and deep-freezes the stored record. Replacing a callback property on the caller-owned definition later does not change dispatch. get()/visible() expose only that frozen snapshot, while schemas() produces detached projections, so the model-visible and executable views cannot drift through a leaked mutable registry object.

Execution: extensible waterfalls plus monotonic policy

ctx.tools.execute() accepts a caller-owned ToolExecutionInput, snapshots it into a pipeline-owned ToolExecution, and runs that call through tools/pre-execute (the reorderable allow/deny/ask waterfall) → registered monotonic guards → tools/execute (around-dispatch wrappers) → tools/post-execute (inspect/replace the result) → tools/result (the immutable authoritative outcome). The outcome is a ToolExecutionResult.

interface ToolExecutionToken {
  readonly [toolExecutionTokenBrand]: true
}

interface ToolExecutionInput {
  readonly callId: CallId
  readonly name: string
  /** Parsed JSON arguments (unknown — tools validate their own input). */
  readonly arguments: unknown
  /** The agent on whose behalf the call runs (set by the agent loop). */
  readonly agent?: Agent
  /**
   * Opaque token of the enclosing transport execution, when one exists. Code
   * Mode sets this on SDK sub-dispatches so commit-style observers can wait for
   * the outer `run_code` outcome without receiving its live mutable execution.
   */
  readonly parent?: ToolExecutionToken
  signal?: AbortSignal
}

interface ToolExecution extends ToolExecutionInput {
  /** Registry-assigned identity shared with nested calls only as their opaque `parent` token. */
  readonly token: ToolExecutionToken
}

ToolExecutionToken is a compile-time opaque type and a frozen, property-free object at runtime; identity comparison is its only operation. Before policy runs, ctx.tools.execute() reads each caller-owned field once, materializes arguments as detached lossless JSON in one recursive pass, assigns a fresh token, and deep-freezes the accepted arguments. A mutable exotic such as Map is rejected and normalized to an error before policy; one-pass materialization prevents a stateful getter from supplying different values to validation and storage. token, callId, name, arguments, agent, and the optional parent token are non-writable throughout all waterfalls, so a listener cannot change which tool or scope the pipeline accepted or reach a live enclosing execution; an around-dispatch wrapper may add, replace, or remove only optional signal. After the complete pipeline the registry freezes the execution and exposes its stable identity to tools/result observers, where the execution remains usable as a WeakMap key without mutation races.

A ToolGuard is scope-aware final pre-dispatch policy. Its shape deliberately has no allow result: undefined preserves the waterfall decision, while a returned reason can only reduce permission, so a later listener cannot undo it.

type ToolGuard = (execution: Readonly<ToolExecution>) => string | undefined

interface ToolExecutionResult {
  callId: CallId
  content: ContentBlock[]
  isError: boolean
  /**
   * Set when the call failed with a {@link HarnessError}: machine-routable
   * `{ name, code }` for retry/sandbox plugins and replay. The model-facing
   * text in `content` is always present; this is extra structure for code.
   */
  error?: ToolErrorInfo
  /**
   * Extra model-facing context a `tools/post-execute` listener attached for the
   * NEXT request (Claude Code's PostToolUse `additionalContext`). It is NOT part
   * of this call's `content` — `content`/`feedback` shape the tool RESULT, but
   * `additionalContext` is a SEPARATE `context/message`. A step can carry
   * multiple tool calls, so the loop BUFFERS every call's `additionalContext`
   * and appends them only AFTER all `tool/result`s for the step, keeping
   * tool-call/result adjacency intact. Carried on the result purely to ferry it
   * from `execute()` up to the loop's per-step buffer.
   */
  additionalContext?: HookContext
  /**
   * The tool-private presentation payload from a successful `execute` (the object
   * return form). Threaded onto the `tool/result` session event and back into
   * {@link ToolResult} for `presentResult`. Opaque (`unknown`); absent when the
   * tool attached none or the call failed.
   */
  meta?: unknown
}

The registry rebuilds and validates the complete authoritative result after post-policy. Its content, structured error, additional context, and presentation metadata must round-trip losslessly through JSON; a malformed or non-JSON value becomes a JSON-safe isError result before tools/result observers run, so the live outcome is always safe for the later durable tool/result append.

Each interception waterfall returns a typed Decision (the idiom shared with the agent/* seams). tools/pre-execute listeners receive (exec, next) and return a PreToolDecision; tools/execute wrappers return a ToolExecutionResult; tools/post-execute listeners receive (exec, result, next) and return a PostToolDecision:

type PreToolDecision =
  | { kind: 'allow' }
  | { kind: 'deny'; reason: string }
  | { kind: 'ask'; reason?: string }

type PostToolDecision =
  | { kind: 'accept'; content?: ContentBlock[]; additionalContext?: HookContext }
  | { kind: 'block'; feedback: ContentBlock[]; additionalContext?: HookContext }

Call next() to delegate to the default (allow / dispatch / accept-unchanged), or return a decision/result to short-circuit. A pre-execute deny skips dispatch and yields an isError result. An ask resolves through the optional approval seam: only allowed-once proceeds, while every non-grant, missing channel/service, or agent-less request becomes a normalized denial. A registered ToolGuard then runs and can still impose a final denial. Input rewrite is deliberately NOT offered on PreToolDecision because it would desync the pre-execution audit/history/UI from what ran. A post-execute accept may replace the model-facing content; a block turns the call into an isError whose content is the corrective feedback. The awaited tools/result notification then receives the frozen execution identity and a deep-frozen result snapshot after every wrapper, post decision, and outer error catch; observers cannot transform the outcome or race each other through payload mutation, and one observer failure neither changes the result nor starves peers. An unregistered tool routes through the same catch as a tool-thrown error, so both failure classes get a structured { name, code } (ToolNotFoundError → UNKNOWN_TOOL) — the loop records a failed tool call instead of failing the whole turn.

The structured-output schema subset

The vocabulary a caller uses to demand a machine-readable result from a subagent (SubagentStartRequest.outputSchema, subagent.md) or a workflow agent() call. It is deliberately NOT full JSON Schema: the schema travels verbatim to the model as a forced tool's parameters, and the produced value is validated client-side by validateStructuredValue — so every accepted keyword must be one the validator actually enforces, and assertSupportedOutputSchema rejects anything else loud (OutputSchemaError, listing every violation). Both walkers reason over own enumerable properties only (JSON carries nothing else) and reject non-plain objects (Date, Map) that would serialize lossily.

type StructuredScalar = string | number | boolean | null

type StructuredSchemaType = 'object' | 'array' | 'string' | 'number' | 'integer' | 'boolean' | 'null'

interface StructuredSchemaNode {
  type: StructuredSchemaType
  properties?: Record<string, StructuredSchemaNode>
  required?: string[]
  additionalProperties?: boolean
  items?: StructuredSchemaNode
  enum?: StructuredScalar[]
  const?: StructuredScalar
  description?: string
  title?: string
  default?: unknown
  examples?: unknown
}

A schema is an object-rooted node (enum/const are scalar-only; description/title/default/examples are annotations, allowed and ignored but still required to be JSON data — they ride the wire):

type StructuredOutputSchema = StructuredSchemaNode & { type: 'object' }

Tool-presentation UI vocabulary

How a tool wants its call shown in a UI (an editor tool-call card, a CLI log line), provider-neutral so a tool describes itself without depending on any client protocol. presentCall/presentResult return a card-tagged render intent — a discriminated union a UI bridge switches on:

• ToolCallView (pending): { card: 'generic', title, kind?, rawInput?, content?, locations? } (the default card; locations is { path, line? }[] files the call reads/modifies, for editor follow-along), { card: 'terminal', title, description?, cwd? } (a shell command → a terminal card), or { card: 'diff', title, diffs, locations? } (a file create/modify → an inline diff card; diffs is { path, oldText, newText }[], oldText: null for a new file).
• ToolResultView (completed): { card: 'generic', title?, content? }, { card: 'terminal', title?, output?, exitCode?, signal? } (the captured run output + exit; a capable UI shows an exit-status pill, an incapable one gets a fenced ```console fallback the bridge derives from output), or { card: 'diff', title?, diffs } (a completed file mutation → the change to show, typically the applied hunks with context lines computed from the before/after content, or a whole-file diff when there is no before-image — e.g. a file create. A tool_call_update's content REPLACES the call's content, so a mutation tool returns this even when it duplicates the call-time snippet, to keep the result from clobbering the diff with result text).

ToolCallKind ('read' | 'edit' | 'delete' | 'move' | 'search' | 'execute' | 'fetch' | 'other') picks an icon on a generic card. FileLocation ({ path, line? }) and FileDiff ({ path, oldText, newText }) are the shared file-card vocabulary. The design is pinned in the render-intent-union RFC; the ACP bridge maps a diff card to a { type: 'diff' } content block, a terminal card to the _meta terminal convention, and relativizes a file card's title against the session cwd.

The full presentation field docs live in packages/core/tools/src/presentation.ts. The bash tool's own schemas (bash/bash_output/bash_kill) and the executor they drive are on bash.md.