feat(llm): add model-specific reasoning effort controls

packages/core/agent-loop/README.md

@@ -60,6 +60,8 @@ The driver owns one agent for its lifetime and runs inside `ctx.agents.withIniti
 
 Every provider call that reaches a successful finish appends exactly one `assistant/message` completion anchor, including content-less calls and `max-tokens` finishes. A successful `agent/step-result` stores its transformed content; a rejected result records empty content before the original failure continues. The anchor retains exact chunk provenance (`[]` for a stream with no chunks) and usage when available, while empty content stays out of derived message history.
 
+After `agent/request` returns a provider/model call config, the loop asks `ctx.llm.resolveCallConfig()` to validate any adapter-owned reasoning effort and materialize its configured default. The effective config is logged in the full `request/header` before dispatch, so a listener can change effort between steps without hidden request drift. A route with no registered adapter preserves the proposed config so an `llm/stream` listener can own and short-circuit it; unhandled terminal dispatch still fails with `NO_ADAPTER`. A new loop instance restores the last effort only when its initial provider/model route exactly matches the logged route; a route change discards that opaque model-owned ID and resolves the new model independently.
+
 Plugin failure ends the current turn, not the loop. Only final adapter dispatch/iteration failures and terminal in-band error or aborted finishes enter `agent/request-error`; middleware, result processing, tools, and `agent/post-step` remain ordinary turn failures. Recovery receives the exact live error, immutable provider facts, and immutable prior failures after the failed step closes. A retry rebuilds from the durable log in a new numbered step, success clears the consecutive history, and exhaustion records the structured failure once on `turn/end`. AgentLoop privately owns one cancellation holder whose explicit signal spans prompt policy, assembly, every step, model and tool work, recovery, continuation, and terminal stop; it retires the holder immediately before publishing `turn/end`, while the driver may remain `running` through the durability flush. An effective `cancel()` emits the typed runtime-only `user | parent` cause before clearing pending work and cooperatively aborting the holder; notification failures cannot veto cancellation, work queued by a notification observer is cleared, work queued by a later abort observer belongs to the next turn, and idle cancellation emits nothing. Durable `turn/end` remains coarse `aborted`; undispatched model tool calls receive synthetic `tool/call` and `ABORTED_BEFORE_DISPATCH` result pairs. Disposal wins terminal classification, and work that ignores the signal must settle before quiescence. The [explicit-cancellation decision](../../../.agents/notes/implemented/architecture/2026-07-16-explicit-turn-cancellation.md) owns the lifecycle and race contract. Terminal continuation stops remain authoritative through turn close and durability flush.
 
 Within a step, exclusive calls form barriers; parallel-safe calls use a bounded rolling pool and are reclassified before start. Only dispatch/body overlaps. Policy, durable results, and result context remain model-ordered. Abort stops new calls, drains started results, then drains accepted batch context before the turn closes through the normal abort path.

packages/core/agent-loop/src/loop.ts

@@ -621,19 +621,43 @@ async function runStep(
 
   // Seed the first request from agent options and later requests from the logged header;
   // detach and freeze so listeners must return an attributable replacement.
-  const seedConfig: LlmCallConfig = deepFreeze(structuredClone(transmission.loggedHeader
-    // eslint-disable-next-line @typescript-eslint/no-non-null-assertion -- loggedHeader ⟹ a snapshot is in the log
-    ? session.requestHeader()!.config
-    : { provider: options.provider ?? '', model: options.model ?? '' }))
+  const loggedConfig = session.requestHeader()?.config
+  const initialProvider = options.provider ?? ''
+  const initialModel = options.model ?? ''
+  const initialConfig: LlmCallConfig = {
+    provider: initialProvider,
+    model: initialModel,
+    ...loggedConfig?.provider === initialProvider
+      && loggedConfig.model === initialModel
+      && loggedConfig.reasoningEffort !== undefined
+      ? { reasoningEffort: loggedConfig.reasoningEffort }
+      : {},
+  }
+  const seedConfig: LlmCallConfig = deepFreeze(structuredClone(
+    transmission.loggedHeader
+      // eslint-disable-next-line @typescript-eslint/no-non-null-assertion -- loggedHeader ⟹ a snapshot is in the log
+      ? session.requestHeader()!.config
+      : initialConfig,
+  ))
 
   // Listener replacements are recorded in the request header before dispatch.
-  const config = await events.waterfall(
+  const proposedConfig = await events.waterfall(
     'agent/request', turn, step, seedConfig, signal, () => Promise.resolve(seedConfig),
   )
   interruptionCheckpoint(signal)
-  if (!config.provider || !config.model) {
+  if (!proposedConfig.provider || !proposedConfig.model) {
     throw new Error(`agent "${agent.id}" has no provider/model: set AgentOptions.provider and AgentOptions.model or supply both via the agent/request waterfall`)
   }
+  let config: LlmCallConfig
+  try {
+    config = await ctx.llm.resolveCallConfig(proposedConfig)
+  } catch (error: unknown) {
+    // A waterfall listener may own and short-circuit a route with no adapter.
+    // Terminal dispatch still raises NO_ADAPTER when no listener handles it.
+    if (!(error instanceof LlmError) || error.code !== 'NO_ADAPTER') throw error
+    config = proposedConfig
+  }
+  interruptionCheckpoint(signal)
 
   // eslint-disable-next-line @typescript-eslint/no-non-null-assertion -- runTurn composes the prefix before every runStep call
   const sessionPrefix = transmission.sessionPrefix!
@@ -651,6 +675,9 @@ async function runStep(
   const request: GenerateOptions = markAgentLoopRequest(deepFreeze({
     provider: header.config.provider,
     model: header.config.model,
+    ...header.config.reasoningEffort !== undefined
+      ? { reasoningEffort: header.config.reasoningEffort }
+      : {},
     messages: [...header.messagePrefix ?? [], ...boundaryMessages],
     ...header.system !== undefined ? { system: header.system } : {},
     ...header.tools !== undefined ? { tools: header.tools } : {},

packages/core/agent-loop/tests/mock-adapter.ts

@@ -1,4 +1,4 @@
-import type { GenerateOptions, StreamChunk } from '@deepseek-ai/dsh-llm'
+import type { GenerateOptions, LlmModelReasoningInfo, StreamChunk } from '@deepseek-ai/dsh-llm'
 import { CallId, LlmAdapter } from '@deepseek-ai/dsh-llm'
 
 /** Helpers to write scripted responses tersely. */
@@ -64,10 +64,20 @@ export function toolCallResponse(rawCallId: string, name: string, args: object,
 export class MockAdapter extends LlmAdapter {
   requests: GenerateOptions[] = []
 
-  constructor(private script: (StreamChunk[] | ((options: GenerateOptions) => StreamChunk[]) | 'hang')[]) {
+  constructor(
+    private script: (StreamChunk[] | ((options: GenerateOptions) => StreamChunk[]) | 'hang')[],
+    private readonly reasoning?: LlmModelReasoningInfo,
+  ) {
     super()
   }
 
+  override resolveModelReasoning(
+    _provider: string,
+    _model: string,
+  ): Promise<LlmModelReasoningInfo | undefined> {
+    return Promise.resolve(this.reasoning)
+  }
+
   async * stream(options: GenerateOptions): AsyncIterable<StreamChunk> {
     this.requests.push(options)
     const entry = this.script.shift()

packages/core/agent-loop/tests/request-reconstruction.spec.ts

@@ -7,7 +7,7 @@
 
 import { describe, expect, it } from 'vitest'
 import { Context } from 'cordis'
-import LlmService from '@deepseek-ai/dsh-llm'
+import LlmService, { LlmError, ReasoningEffortId } from '@deepseek-ai/dsh-llm'
 import type { GenerateOptions } from '@deepseek-ai/dsh-llm'
 import SessionStore, { Session, SessionId, foldRequestHeader } from '@deepseek-ai/dsh-session'
 import SystemPrompt from '@deepseek-ai/dsh-system-prompt'
@@ -104,6 +104,81 @@ describe('request stability across the loop', () => {
     expectPrefixExtension(adapter.requests[0]!, adapter.requests[1]!)
   })
 
+  it('logs adapter defaults, supports per-turn effort changes, and restores the effective value', async () => {
+    const reasoning = {
+      efforts: [
+        { id: ReasoningEffortId('high'), name: 'High' },
+        { id: ReasoningEffortId('max'), name: 'Max' },
+      ],
+      defaultEffort: ReasoningEffortId('high'),
+    }
+    const adapter = new MockAdapter([textResponse('one'), textResponse('two')], reasoning)
+    const ctx = await harness(adapter)
+    const agent = ctx.agentLoop.create(SessionId('effort'), { provider: 'mock', model: 'mock' })
+    ctx.on('agent/request', async (_agent, turn, _step, _config, _signal, next) => {
+      const config = await next()
+      return turn === 2 ? { ...config, reasoningEffort: ReasoningEffortId('max') } : config
+    })
+
+    send(agent, 'first')
+    await waitForIdle(ctx, agent)
+    send(agent, 'second')
+    await waitForIdle(ctx, agent)
+
+    expect(adapter.requests.map(request => request.reasoningEffort)).toEqual([
+      ReasoningEffortId('high'),
+      ReasoningEffortId('max'),
+    ])
+    const headers = agent.session.events.filter(event => event.type === 'request/header')
+    expect(headers.map(event => event.data.header.config.reasoningEffort)).toEqual([
+      ReasoningEffortId('high'),
+      ReasoningEffortId('max'),
+    ])
+    expect(headers.map(event => event.data.reason)).toEqual(['initial', 'change'])
+
+    const resumedAdapter = new MockAdapter([textResponse('three')], reasoning)
+    const resumedCtx = await harness(resumedAdapter)
+    const resumedHandle = await resumedCtx.agents.create({
+      sessionId: SessionId('effort-resumed'),
+      seed: structuredClone(agent.session.events),
+      agentOptions: { provider: 'mock', model: 'mock' },
+    })
+    send(resumedHandle.agent, 'third')
+    await waitForIdle(resumedCtx, resumedHandle.agent)
+
+    expect(resumedAdapter.requests[0]?.reasoningEffort).toBe(ReasoningEffortId('max'))
+    const resumedHeaders = resumedHandle.agent.session.events.filter(event => event.type === 'request/header')
+    expect(resumedHeaders.at(-1)?.data.header.config.reasoningEffort).toBe(ReasoningEffortId('max'))
+    expect(resumedHeaders.at(-1)?.data.reason).toBe('resume')
+  })
+
+  it.each(['plain error', 'LLM error'] as const)(
+    'does not swallow a %s from reasoning resolution',
+    async (kind) => {
+      const failure = kind === 'plain error'
+        ? new Error('reasoning metadata failed')
+        : new LlmError('unsupported effort', 'UNSUPPORTED_REASONING_EFFORT')
+      const adapter = new class extends MockAdapter {
+        override resolveModelReasoning(): Promise<never> {
+          return Promise.reject(failure)
+        }
+      }([])
+      const ctx = await harness(adapter)
+      const errors: Error[] = []
+      ctx.on('agent/error', (_agent, _turn, _step, error) => void errors.push(error))
+      const agent = ctx.agentLoop.create(SessionId(`reasoning-${kind}`), {
+        provider: 'mock',
+        model: 'mock',
+      })
+
+      send(agent, 'go')
+      await waitForIdle(ctx, agent)
+
+      expect(errors).toContain(failure)
+      expect(adapter.requests).toHaveLength(0)
+    },
+  )
+
   it('a compaction replace rewrites the resend, and the log explains it', async () => {
     const adapter = new MockAdapter([textResponse('one'), textResponse('two')])
     const ctx = await harness(adapter)
@@ -301,6 +376,7 @@ describe('request stability across the loop', () => {
       const firstChunk = events.find(e => e.type === 'assistant/chunk' && e.seq > stepStart.seq)!
       const header = foldRequestHeader(events.slice(0, firstChunk.seq))!
       expect(request.model).toBe(header.config.model)
+      expect(request.reasoningEffort).toBe(header.config.reasoningEffort)
       expect(request.system).toEqual(header.system)
       expect(structuredClone(request.tools ?? [])).toEqual(structuredClone(header.tools ?? []))
       expect(request.temperature).toBe(header.config.temperature)