Merge branch 'structured-output-subagent-seam' into worktree-dynamic-workflows
This commit is contained in:
@@ -119,9 +119,14 @@ export function startInProcessRun(
|
||||
if (request.maxDepth !== undefined && childDepth > request.maxDepth) {
|
||||
throw new SubagentDepthError(childDepth, request.maxDepth)
|
||||
}
|
||||
// Assert the schema subset BEFORE any child exists (the service has already
|
||||
// capability-gated; this rejects a schema outside the enforced subset loud).
|
||||
const schema = request.outputSchema
|
||||
// Snapshot, then assert, the schema subset BEFORE any child exists (the
|
||||
// service has already capability-gated; this rejects a schema outside the
|
||||
// enforced subset loud). The snapshot is load-bearing: the caller keeps its
|
||||
// reference, so validating and attaching the ORIGINAL would let a
|
||||
// post-start() mutation drift the enforced schema away from the asserted
|
||||
// one — the clone pins assertion, the model-visible parameters, and
|
||||
// validateStructuredValue to the same isolation-immutable value.
|
||||
const schema = request.outputSchema === undefined ? undefined : structuredClone(request.outputSchema)
|
||||
if (schema !== undefined) assertSupportedOutputSchema(schema)
|
||||
|
||||
const childId = AgentId(randomUUID())
|
||||
|
||||
@@ -25,7 +25,11 @@
|
||||
* output is captured — without it, the loop's default "had tool calls ⇒
|
||||
* continue" buys a wasted extra model step per structured child. It is also
|
||||
* `prepend: true`: the veto must run before any earlier-registered listener
|
||||
* that could short-circuit the chain into a forced continue.
|
||||
* that could short-circuit the chain into a forced continue. A third listener
|
||||
* closes the within-step window the continuation veto cannot: a
|
||||
* `tools/pre-execute` deny for any call arriving after the agent's capture, so
|
||||
* a response that lists `structured_output` before further tool calls cannot
|
||||
* run side effects after the final answer was accepted.
|
||||
*
|
||||
* Lifetime is refcounted with two kinds of holder: each backend acquires for
|
||||
* its plugin lifetime (so the tool exists before any run), and each structured
|
||||
@@ -42,7 +46,7 @@ import type { Agent } from '@deepseek-ai/dsh-agent'
|
||||
import type { ContentBlock, ToolSchema } from '@deepseek-ai/dsh-llm'
|
||||
import type { ContinuationDecision } from '@deepseek-ai/dsh-agent'
|
||||
import type { AssembleContext, PromptAssembly } from '@deepseek-ai/dsh-system-prompt'
|
||||
import type { ToolExecution } from '@deepseek-ai/dsh-tools'
|
||||
import type { PreToolDecision, ToolExecution } from '@deepseek-ai/dsh-tools'
|
||||
import { ToolArgsError, validateStructuredValue, type StructuredOutputSchema } from '@deepseek-ai/dsh-tools'
|
||||
|
||||
/** The model-facing tool name a structured child must call to finish. */
|
||||
@@ -236,4 +240,27 @@ function registerRuntime(root: Context, runtime: StructuredRuntime): void {
|
||||
if (runtime.states.get(agent)?.captured) return Promise.resolve({ action: 'stop' })
|
||||
return next()
|
||||
}, { prepend: true }))
|
||||
|
||||
// Terminal means terminal WITHIN the step, not only at its end: the
|
||||
// turn-continuation veto above runs after every call in the current model
|
||||
// response has executed, so a response that puts `structured_output` before
|
||||
// further tool calls would still perform those side effects after the final
|
||||
// answer was accepted. Deny every later call for a captured agent at the
|
||||
// allow/deny gate — dispatch is skipped and the model sees an `isError`
|
||||
// result naming the contract. Calls that PRECEDE the capture in the same
|
||||
// response ran before `captured` was set and are untouched; a second
|
||||
// `structured_output` is denied like any other call. `prepend: true` for the
|
||||
// same reason as the continuation veto: no earlier-registered allow may
|
||||
// short-circuit past the terminal contract.
|
||||
runtime.disposers.push(root.on('tools/pre-execute', function (
|
||||
this: unknown, exec: ToolExecution, next: () => Promise<PreToolDecision>,
|
||||
): Promise<PreToolDecision> {
|
||||
if (exec.agent && runtime.states.get(exec.agent)?.captured) {
|
||||
return Promise.resolve({
|
||||
kind: 'deny',
|
||||
reason: `structured output already recorded: the run is complete, so \`${exec.name}\` is not executed`,
|
||||
})
|
||||
}
|
||||
return next()
|
||||
}, { prepend: true }))
|
||||
}
|
||||
@@ -1,6 +1,6 @@
|
||||
import { describe, expect, it } from 'vitest'
|
||||
import { Context } from 'cordis'
|
||||
import LlmService, { type GenerateOptions } from '@deepseek-ai/dsh-llm'
|
||||
import LlmService, { CallId, type ContentBlock, type GenerateOptions } from '@deepseek-ai/dsh-llm'
|
||||
import SessionStore from '@deepseek-ai/dsh-session'
|
||||
import SystemPrompt from '@deepseek-ai/dsh-system-prompt'
|
||||
import ToolRegistry from '@deepseek-ai/dsh-tools'
|
||||
@@ -86,6 +86,90 @@ describe('in-process structured output', () => {
|
||||
await run.dispose()
|
||||
})
|
||||
|
||||
it('denies tool calls that FOLLOW the capture in the same response — terminal means terminal', async () => {
|
||||
// One model response carrying structured_output FIRST and a side-effecting
|
||||
// call after it: the continuation veto only fires at step end, so without
|
||||
// the pre-execute deny the trailing call would still run after the final
|
||||
// answer was accepted.
|
||||
const response = [
|
||||
...toolCallResponse('c1', STRUCTURED_OUTPUT_TOOL, { answer: 5 }).slice(0, -2),
|
||||
{ type: 'block-start', index: 1, blockType: 'tool-call' },
|
||||
{ type: 'block-end', index: 1, block: { type: 'tool-call', id: CallId('c2'), name: 'side_effect', arguments: '{}' } },
|
||||
{ type: 'usage', usage: { inputTokens: 10, outputTokens: 5 } },
|
||||
{ type: 'finish', reason: { kind: 'tool-calls' } },
|
||||
] as Script[number]
|
||||
const { ctx, parent } = await setup([response])
|
||||
let sideEffectRan = false
|
||||
ctx.tools.register({
|
||||
name: 'side_effect',
|
||||
description: 'probe',
|
||||
parameters: { type: 'object', properties: {} },
|
||||
execute(): Promise<ContentBlock[]> {
|
||||
sideEffectRan = true
|
||||
return Promise.resolve([{ type: 'text', text: 'ran' }])
|
||||
},
|
||||
})
|
||||
const run = ctx.subagents.start('spawn', structuredRequest(parent))
|
||||
const result = await run.result
|
||||
expect(result.stopReason).toBe('completed')
|
||||
expect(result.structured).toEqual({ answer: 5 })
|
||||
// The deny skipped dispatch entirely: the probe body never ran.
|
||||
expect(sideEffectRan).toBe(false)
|
||||
await run.dispose()
|
||||
})
|
||||
|
||||
it('leaves tool calls that PRECEDE the capture in the same response untouched', async () => {
|
||||
const response = [
|
||||
{ type: 'block-start', index: 0, blockType: 'tool-call' },
|
||||
{ type: 'block-end', index: 0, block: { type: 'tool-call', id: CallId('c1'), name: 'side_effect', arguments: '{}' } },
|
||||
...toolCallResponse('c2', STRUCTURED_OUTPUT_TOOL, { answer: 6 }).map(chunk =>
|
||||
'index' in chunk ? { ...chunk, index: 1 } : chunk),
|
||||
] as Script[number]
|
||||
const { ctx, parent } = await setup([response])
|
||||
let sideEffectRan = false
|
||||
ctx.tools.register({
|
||||
name: 'side_effect',
|
||||
description: 'probe',
|
||||
parameters: { type: 'object', properties: {} },
|
||||
execute(): Promise<ContentBlock[]> {
|
||||
sideEffectRan = true
|
||||
return Promise.resolve([{ type: 'text', text: 'ran' }])
|
||||
},
|
||||
})
|
||||
const run = ctx.subagents.start('spawn', structuredRequest(parent))
|
||||
const result = await run.result
|
||||
// The call ran BEFORE captured was set: the deny gate only guards the
|
||||
// window after the terminal answer landed.
|
||||
expect(sideEffectRan).toBe(true)
|
||||
expect(result.structured).toEqual({ answer: 6 })
|
||||
await run.dispose()
|
||||
})
|
||||
|
||||
it('snapshots the schema at start(): caller mutation after start cannot drift enforcement', async () => {
|
||||
const mutable: StructuredOutputSchema = {
|
||||
type: 'object',
|
||||
properties: { answer: { type: 'number' } },
|
||||
required: ['answer'],
|
||||
additionalProperties: false,
|
||||
}
|
||||
const pristine = structuredClone(mutable)
|
||||
const { ctx, parent, adapter } = await setup([
|
||||
toolCallResponse('c1', STRUCTURED_OUTPUT_TOOL, { answer: 3 }),
|
||||
])
|
||||
const run = ctx.subagents.start('spawn', structuredRequest(parent, { outputSchema: mutable }))
|
||||
// Mutate the caller's object AFTER start() returned but before the child's
|
||||
// first request assembles: with a live reference this would reach both the
|
||||
// model-visible parameters and validateStructuredValue.
|
||||
;(mutable.properties as Record<string, unknown>).answer = { type: 'string' }
|
||||
const result = await run.result
|
||||
expect(result.structured).toEqual({ answer: 3 })
|
||||
// The child's request carried the PRISTINE schema, not the mutated one.
|
||||
const childRequest = adapter.requests.at(-1)
|
||||
const captureTool = (childRequest?.tools ?? []).find(tool => tool.name === STRUCTURED_OUTPUT_TOOL)
|
||||
expect(captureTool?.parameters).toEqual(pristine)
|
||||
await run.dispose()
|
||||
})
|
||||
|
||||
it('the captured-turn veto is prepend: an EARLIER force-continue listener cannot short-circuit it', async () => {
|
||||
const ctx = new Context()
|
||||
await ctx.plugin(SystemPrompt)
|
||||
|
||||
Reference in New Issue
Block a user