Coder/deepseek-harness
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

refactor: identify and freeze messages at creation

Browse Source
This commit is contained in:
_Kerman
2026-07-28 13:55:59 +08:00
parent c49c0ba497
commit fbf87e660c
345 changed files with 5220 additions and 2901 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
packages/subagent/subagent-acp/tests/loader-composition.e2e.ts
View File

@@ -64,7 +64,7 @@ describe('ACP subagent cwd inheritance through a real cordis.yml', () => {
// session's workspace, never the harness process's launch directory.
const results = events.filter(event => event.type === 'tool/result')
expect(results).toHaveLength(1)
const resultText = results[0]!.data.content
const resultText = results[0]!.data.message.content[0].content
.filter(block => block.type === 'text')
.map(block => block.text)
.join('')

2
packages/subagent/subagent-dsh-sdk/src/run.ts
View File

@@ -170,7 +170,7 @@ export async function startSdkRun(request: SubagentStartRequest, spec: SdkRunSpe
if (event.type === 'assistant/chunk' && event.data.chunk.type === 'text-delta') {
partial.push(event.data.chunk.text)
} else if (event.type === 'assistant/message') {
lastMessage = event.data.content
lastMessage = event.data.message.content
}
}
const collectOutput = (): ContentBlock[] => {

2
packages/subagent/subagent-dsh-sdk/tests/loader-composition.e2e.ts
View File

@@ -91,7 +91,7 @@ describe('SDK subagent cwd inheritance through a real cordis.yml', () => {
// process's launch directory.
const results = events.filter(event => event.type === 'tool/result')
expect(results).toHaveLength(1)
const resultText = results[0]!.data.content
const resultText = results[0]!.data.message.content[0].content
.filter(block => block.type === 'text')
.map(block => block.text)
.join('')

7
packages/subagent/subagent-fork/tests/multi-subagent.spec.ts
View File

@@ -1,3 +1,4 @@
import { createUserMessage } from '@deepseek-ai/dsh-llm'
import { describe, expect, it } from 'vitest'
import { Context } from 'cordis'
import { SessionId } from '@deepseek-ai/dsh-session'
@@ -64,7 +65,7 @@ describe('multi-subagent coexistence (spawn + fork on one context)', () => {
])
// Parent does one real turn first, so the fork has a completed turn to seed.
parent.followup({ content: [{ type: 'text', text: 'parent q1' }], source: { kind: 'user' } })
parent.followup(createUserMessage({ content: [{ type: 'text', text: 'parent q1' }], source: { kind: 'user' } }))
await parent.whenIdle()
const parentPrefixLen = parent.session.events.length
@@ -93,10 +94,10 @@ describe('multi-subagent coexistence (spawn + fork on one context)', () => {
await forkRun.dispose()
// The parent is unaffected and keeps working after both delegations.
parent.followup({ content: [{ type: 'text', text: 'parent q2' }], source: { kind: 'user' } })
parent.followup(createUserMessage({ content: [{ type: 'text', text: 'parent q2' }], source: { kind: 'user' } }))
await parent.whenIdle()
const lastParentMessage = parent.session.events.findLast(e => e.type === 'assistant/message')
expect(lastParentMessage?.type === 'assistant/message' && text(lastParentMessage.data.content)).toBe('parent turn two')
expect(lastParentMessage?.type === 'assistant/message' && text(lastParentMessage.data.message.content)).toBe('parent turn two')
// The parent's OWN log never recorded the children's internal steps — its
// only subagent-related entries would be tool/call+tool/result IF it had
// used the tool, but here we called the service directly, so the parent log

15
packages/subagent/subagent-fork/tests/subagent-fork.spec.ts
View File

@@ -1,3 +1,4 @@
import { createUserMessage } from '@deepseek-ai/dsh-llm'
import { describe, expect, it } from 'vitest'
import { Context } from 'cordis'
import Loader from '@cordisjs/plugin-loader'
@@ -89,9 +90,9 @@ describe('dsh-subagent-fork', () => {
it('seeds every completed parent turn through the last turn/end', async () => {
const { ctx, parent } = await setup([textResponse('first'), textResponse('second'), textResponse('child')])
parent.followup({ content: [{ type: 'text', text: 'q1' }], source: { kind: 'user' } })
parent.followup(createUserMessage({ content: [{ type: 'text', text: 'q1' }], source: { kind: 'user' } }))
await parent.whenIdle()
parent.followup({ content: [{ type: 'text', text: 'q2' }], source: { kind: 'user' } })
parent.followup(createUserMessage({ content: [{ type: 'text', text: 'q2' }], source: { kind: 'user' } }))
await parent.whenIdle()
const parentPrefixLen = parent.session.events.length
@@ -108,7 +109,7 @@ describe('dsh-subagent-fork', () => {
// Parent runs one turn, then we fork. The child's seeded log should contain
// the parent's first turn, and the child should run its own new turn on top.
const { ctx, parent } = await setup([textResponse('parent answer'), textResponse('child answer')])
parent.followup({ content: [{ type: 'text', text: 'parent question' }], source: { kind: 'user' } })
parent.followup(createUserMessage({ content: [{ type: 'text', text: 'parent question' }], source: { kind: 'user' } }))
await parent.whenIdle()
const parentPrefixLen = parent.session.events.length
@@ -137,10 +138,10 @@ describe('dsh-subagent-fork', () => {
// open (a hanging model call), and fork while it's in flight. The seed must stop after the
// balanced first turn; including the open turn would fail invariant replay during start.
const { ctx, parent } = await setup([textResponse('done'), 'hang', textResponse('child')])
parent.followup({ content: [{ type: 'text', text: 'q1' }], source: { kind: 'user' } })
parent.followup(createUserMessage({ content: [{ type: 'text', text: 'q1' }], source: { kind: 'user' } }))
await parent.whenIdle()
// Start a second turn that hangs (open turn/start + open step, never ends).
parent.followup({ content: [{ type: 'text', text: 'q2' }], source: { kind: 'user' } })
parent.followup(createUserMessage({ content: [{ type: 'text', text: 'q2' }], source: { kind: 'user' } }))
await new Promise(r => setTimeout(r, 20)) // let the hanging turn open
// Forking now must NOT throw (the open second turn is excluded from the seed).
@@ -164,7 +165,7 @@ describe('dsh-subagent-fork', () => {
textResponse('parent turn'),
toolCallResponse('c1', STRUCTURED_OUTPUT_TOOL, { answer: 9 }),
])
parent.followup({ content: [{ type: 'text', text: 'warm up' }], source: { kind: 'user' } })
parent.followup(createUserMessage({ content: [{ type: 'text', text: 'warm up' }], source: { kind: 'user' } }))
await parent.whenIdle()
const run = await start(ctx, 'fork', {
prompt: [{ type: 'text', text: 'report structured' }],
@@ -183,7 +184,7 @@ describe('dsh-subagent-fork', () => {
// `readResult` must scan only child-owned events after the seed. The child emits no assistant
// message, so scanning the whole log would incorrectly return the parent's distinctive text.
const { ctx, parent } = await setup([textResponse('parent stale'), emptyStop])
parent.followup({ content: [{ type: 'text', text: 'parent question' }], source: { kind: 'user' } })
parent.followup(createUserMessage({ content: [{ type: 'text', text: 'parent question' }], source: { kind: 'user' } }))
await parent.whenIdle()
const run = await start(ctx, 'fork', { prompt: [{ type: 'text', text: 'child question' }], parent })

6
packages/subagent/subagent-inprocess/src/index.ts
View File

@@ -11,7 +11,7 @@ import { randomUUID } from 'node:crypto'
import type { Context } from 'cordis'
import type { Agent, AgentOptions } from '@deepseek-ai/dsh-agent'
import { findLastMessageTurnEnd, SessionId, type SessionEvent, type TurnEndReason } from '@deepseek-ai/dsh-session'
import type { ContentBlock } from '@deepseek-ai/dsh-llm'
import { createUserMessage, type ContentBlock } from '@deepseek-ai/dsh-llm'
import { assertSubagentMaxDepth, delegationDepthOf } from '@deepseek-ai/dsh-subagent'
import type { SubagentResult, SubagentRun, SubagentStartRequest, SubagentStopReason } from '@deepseek-ai/dsh-subagent'
import {
@@ -141,7 +141,7 @@ export async function startInProcessRun(
const result: Promise<SubagentResult> = (async () => {
try {
child.followup({ content: request.prompt, source: { kind: 'user' } })
child.followup(createUserMessage({ content: request.prompt, source: { kind: 'user' } }))
await child.whenIdle()
return readResult(
child,
@@ -176,7 +176,7 @@ function readResult(
const own = child.session.events.slice(seedLength)
const lastMessage = own.findLast((event): event is SessionEvent<'assistant/message'> => event.type === 'assistant/message')
const lastEnd = findLastMessageTurnEnd(own)
const output: ContentBlock[] = lastMessage?.data.content ?? []
const output: ContentBlock[] = lastMessage?.data.message.content ?? []
const recorded = toStopReason(lastEnd?.data.reason)
// Disposal can tear the owner down before the loop records its ordinary
// `aborted` end, yielding `disposed` instead. A requested cancellation owns

26
packages/subagent/subagent-inprocess/tests/structured.spec.ts
View File

@@ -1,6 +1,6 @@
import { describe, expect, it } from 'vitest'
import { Context } from 'cordis'
import { CallId, type ContentBlock, type GenerateOptions } from '@deepseek-ai/dsh-llm'
import { createUserMessage, CallId, type ContentBlock, type GenerateOptions } from '@deepseek-ai/dsh-llm'
import { SessionId } from '@deepseek-ai/dsh-session'
import AgentLoop from '@deepseek-ai/dsh-agent-loop'
import { mountAgentLoopTestDependencies } from '@deepseek-ai/dsh-agent-loop-testkit'
@@ -185,8 +185,8 @@ describe('in-process structured output', () => {
expect(sideEffectRan).toBe(false)
const child = ctx.agents.get(run.id)
const sideEffectResult = child?.session.events.find(event =>
event.type === 'tool/result' && event.data.callId === 'c2')
expect(sideEffectResult?.type === 'tool/result' && sideEffectResult.data.isError).toBe(true)
event.type === 'tool/result' && event.data.message.source.callId === 'c2')
expect(sideEffectResult?.type === 'tool/result' && sideEffectResult.data.message.content[0].isError).toBe(true)
await run.dispose()
})
@@ -312,8 +312,8 @@ describe('in-process structured output', () => {
// ...the logged tool result is the blocked isError with the feedback...
const child = ctx.agents.get(run.id)!
const results = child.session.events.filter(e => e.type === 'tool/result')
expect((results[0]!.data as { isError?: boolean }).isError).toBe(true)
expect(JSON.stringify((results[0]!.data as { content: unknown }).content)).toContain('capture rejected by hook')
expect(results[0]!.data.message.content[0].isError).toBe(true)
expect(JSON.stringify(results[0]!.data.message.content)).toContain('capture rejected by hook')
// ...and the turn CONTINUED past the blocked call (no captured veto):
// the model got to react to the failure with a second step.
expect(adapter.requests.length).toBe(2)
@@ -357,8 +357,8 @@ describe('in-process structured output', () => {
expect(result.stopReason).toBe('error')
const child = ctx.agents.get(run.id)
const captureResult = child?.session.events.find(event =>
event.type === 'tool/result' && event.data.callId === 'c1')
expect(captureResult?.type === 'tool/result' && captureResult.data.isError).toBe(true)
event.type === 'tool/result' && event.data.message.source.callId === 'c1')
expect(captureResult?.type === 'tool/result' && captureResult.data.message.content[0].isError).toBe(true)
await run.dispose()
})
@@ -428,8 +428,8 @@ describe('in-process structured output', () => {
expect(adapter.requests).toHaveLength(2)
const child = ctx.agents.get(run.id)!
const outer = child.session.events.find(event =>
event.type === 'tool/result' && event.data.callId === CallId('c1'))
expect(outer?.type === 'tool/result' && outer.data.isError).toBe(true)
event.type === 'tool/result' && event.data.message.source.callId === CallId('c1'))
expect(outer?.type === 'tool/result' && outer.data.message.content[0].isError).toBe(true)
await run.dispose()
})
@@ -463,7 +463,7 @@ describe('in-process structured output', () => {
textResponse('parent answer'),
toolCallResponse('c1', STRUCTURED_OUTPUT_TOOL, { answer: 1 }),
])
parent.followup({ content: [{ type: 'text', text: 'hello' }], source: { kind: 'user' } })
parent.followup(createUserMessage({ content: [{ type: 'text', text: 'hello' }], source: { kind: 'user' } }))
await parent.whenIdle()
expect(adapter.requests[0]!.system ?? '').not.toContain(STRUCTURED_OUTPUT_INSTRUCTION)
const run = await ctx.subagents.start('spawn', structuredRequest(parent))
@@ -479,7 +479,7 @@ describe('in-process structured output', () => {
describe('scoped registration (each child owns its capture tool)', () => {
it('a plain agent never sees the tool: nothing is registered globally at all', async () => {
const { ctx, parent, adapter } = await setup([textResponse('parent answer')])
parent.followup({ content: [{ type: 'text', text: 'hello' }], source: { kind: 'user' } })
parent.followup(createUserMessage({ content: [{ type: 'text', text: 'hello' }], source: { kind: 'user' } }))
await parent.whenIdle()
// Scoped registration: the global view has no capture tool, ever.
expect(ctx.tools.get(STRUCTURED_OUTPUT_TOOL)).toBeUndefined()
@@ -493,7 +493,7 @@ describe('in-process structured output', () => {
// Child turn: must see it, with the run's schema.
toolCallResponse('c1', STRUCTURED_OUTPUT_TOOL, { answer: 42 }),
])
parent.followup({ content: [{ type: 'text', text: 'hello' }], source: { kind: 'user' } })
parent.followup(createUserMessage({ content: [{ type: 'text', text: 'hello' }], source: { kind: 'user' } }))
await parent.whenIdle()
expect(toolNames(adapter.requests[0]!)).not.toContain(STRUCTURED_OUTPUT_TOOL)
@@ -571,7 +571,7 @@ describe('in-process structured output', () => {
it('a non-structured agent request keeps tools ABSENT when it had none (no tools: [] materialized)', async () => {
const { parent, adapter } = await setup([textResponse('plain')])
parent.followup({ content: [{ type: 'text', text: 'q' }], source: { kind: 'user' } })
parent.followup(createUserMessage({ content: [{ type: 'text', text: 'q' }], source: { kind: 'user' } }))
await parent.whenIdle()
const request = adapter.requests[0]!
expect(request.tools).toBeUndefined()

7
packages/subagent/subagent-inprocess/tests/subagent-inprocess.spec.ts
View File

@@ -1,3 +1,4 @@
import { createUserMessage } from '@deepseek-ai/dsh-llm'
import { describe, expect, it } from 'vitest'
import { Context } from 'cordis'
import { type Agent } from '@deepseek-ai/dsh-agent'
@@ -68,10 +69,10 @@ describe('startInProcessRun', () => {
turn,
trigger: { kind: 'injection', source: { kind: 'plugin', plugin: 'late-metadata' } },
})
session.append('user/message', {
session.append('user/message', createUserMessage({
content: [{ type: 'text', text: 'late metadata' }],
source: { kind: 'plugin', plugin: 'late-metadata' },
}, { surfaceOp: 'append' })
}), { surfaceOp: 'append' })
session.append('turn/end', { turn, reason: { kind: 'completed' } })
})
@@ -88,7 +89,7 @@ describe('startInProcessRun', () => {
it('seeds a forked child but reads only the child-owned output', async () => {
const { ctx, parent } = await setup([textResponse('parent answer'), textResponse('child answer')])
parent.followup({ content: [{ type: 'text', text: 'parent question' }], source: { kind: 'user' } })
parent.followup(createUserMessage({ content: [{ type: 'text', text: 'parent question' }], source: { kind: 'user' } }))
await parent.whenIdle()
const seed = parent.session.events.slice()
const run = await startInProcessRun(request(parent), { seed })

6
packages/subagent/subagent-spawn/tests/spawn.e2e.ts
View File

@@ -1,3 +1,4 @@
import { createUserMessage } from '@deepseek-ai/dsh-llm'
import { mkdtemp, readFile, rm } from 'node:fs/promises'
import { tmpdir } from 'node:os'
import { join } from 'node:path'
@@ -24,10 +25,11 @@ describe.skipIf(!process.env.DEEPSEEK_API_KEY)('spawn backend with-key smoke', (
ctx = await spawnHarness(workdir)
const parent = ctx.agentLoop.create(SessionId('e2e-parent'), { provider: 'deepseek', model: 'deepseek-v4-flash' })
parent.followup({ content: [{ type: 'text', text:
parent.followup(createUserMessage({
content: [{ type: 'text', text:
'Use the subagent tool to delegate this exact task: "Use the bash tool to write the text '
+ 'SUBAGENT_WAS_HERE into a file named proof.txt in the current directory." '
+ 'After the subagent finishes, tell me it is done.' }], source: { kind: 'user' } })
+ 'After the subagent finishes, tell me it is done.' }], source: { kind: 'user' } }))
await waitForIdle(ctx, parent)
// Assert the filesystem effect independently of the model response.

5
packages/subagent/subagent-spawn/tests/subagent-spawn.spec.ts
View File

@@ -1,3 +1,4 @@
import { createUserMessage } from '@deepseek-ai/dsh-llm'
import { describe, expect, it } from 'vitest'
import { Context, symbols, type EffectMeta } from 'cordis'
import Loader from '@cordisjs/plugin-loader'
@@ -106,7 +107,7 @@ describe('dsh-subagent-spawn', () => {
it('a fresh child does NOT inherit the parent conversation (its log starts empty before the prompt)', async () => {
// Drive the parent through one real turn so it has history, THEN spawn.
const { ctx, parent } = await setup([textResponse('parent turn'), textResponse('child sees nothing')])
parent.followup({ content: [{ type: 'text', text: 'parent prompt' }], source: { kind: 'user' } })
parent.followup(createUserMessage({ content: [{ type: 'text', text: 'parent prompt' }], source: { kind: 'user' } }))
await parent.whenIdle()
const parentEventCount = parent.session.events.length
expect(parentEventCount).toBeGreaterThan(0)
@@ -383,7 +384,7 @@ describe('dsh-subagent-spawn', () => {
textResponse('parent answer'),
textResponse('child answer'),
])
parent.followup({ content: [{ type: 'text', text: 'hi' }], source: { kind: 'user' } })
parent.followup(createUserMessage({ content: [{ type: 'text', text: 'hi' }], source: { kind: 'user' } }))
await parent.whenIdle()
const run = await start(ctx, 'spawn', {