fix(agent-loop): rematerialize adapter defaults

packages/llm/llm/README.i18n.yaml

@@ -2,5 +2,5 @@
 # side as of the last confirmed-consistent state. Both languages carry equal authority;
 # after editing either side, bring the other along and re-record with:
 #   pnpm run verify-translation-pairing --write packages/llm/llm/README.md
-README.md: 3ffcfb59fa7d3e3077a59d0b0c8ebd9afa590e7e
-README.zh.md: df44b2bad1d5fa5c03dff86de584bfff63dbf297
+README.md: 2dbd530ca17ef34787cb4195c04ca85d768980b7
+README.zh.md: 9928113f5cbfc49887980fde57ad4ee9f37dbd22

packages/llm/llm/README.md

@@ -25,7 +25,7 @@ Provider and model metadata is a discovery surface, not a routing whitelist. `re
 
 Exact-model metadata is a separate correctness query, not a catalog decoration or global LLM setting. `resolveModelInfo()` asks the adapter that owns the exact provider/model route once; an adapter can describe an unlisted dynamic model, and absent `context`, `defaultMaxTokens`, or `reasoning` fields preserve unknown capacity, provider-owned output defaults, or unavailable reasoning capability. Invalid identity, context, output default, or reasoning metadata fails with `INVALID_MODEL_INFO`, `INVALID_MODEL_CONTEXT`, `INVALID_MODEL_MAX_TOKENS`, or `INVALID_MODEL_REASONING`.
 
-`defaultMaxTokens` is an adapter-configured per-request output cap, not a model hard limit. `resolveCallConfig()` materializes it only when the request omits `maxTokens`; an explicit cap wins. Reasoning identifiers are opaque adapter-owned strings rather than a core enum: the same resolution accepts only an exact advertised identifier, materializes `defaultEffort` when present, and otherwise preserves the provider default. Asynchronous model resolvers receive the caller's signal and must settle promptly after cancellation. `prepareCall()` additionally retains the exact adapter registration through header logging and terminal dispatch, so HMR cannot combine one adapter's capability result with another adapter's request; reusing its one-shot handle or changing its call-config fields fails with `INVALID_PREPARED_CALL`. An unsupported explicit or configured effort fails with `UNSUPPORTED_REASONING_EFFORT` before provider I/O.
+`defaultMaxTokens` is an adapter-configured per-request output cap, not a model hard limit. `resolveCallConfig()` materializes it only when the request omits `maxTokens`; an explicit cap wins. Reasoning identifiers are opaque adapter-owned strings rather than a core enum: the same resolution accepts only an exact advertised identifier, materializes `defaultEffort` when present, and otherwise preserves the provider default. Asynchronous model resolvers receive the caller's signal and must settle promptly after cancellation. `prepareCall()` additionally reports which `maxTokens` and `reasoningEffort` fields it materialized in `adapterDefaults` and retains the exact adapter registration through header logging and terminal dispatch, so HMR cannot combine one adapter's capability result with another adapter's request; reusing its one-shot handle or changing its call-config fields fails with `INVALID_PREPARED_CALL`. An unsupported explicit or configured effort fails with `UNSUPPORTED_REASONING_EFFORT` before provider I/O.
 
 ### Events
 
@@ -48,7 +48,7 @@ Streaming is a raw chunk protocol (`block-start`, `text-delta`, `reasoning-delta
 
 ### Call configuration (`call-config.ts`)
 
-`LlmCallConfig` is the provider, model, optional adapter-owned reasoning effort, and sampling scalars of one conversation's requests (`provider`, `model`, `reasoningEffort`, `temperature`, `maxTokens`, `stop` — each mapping 1:1 onto the same-named `GenerateOptions` field). It is per-conversation state recorded in the session log as part of the request header (see the dsh-session `request/header` events), never a silently-adjustable per-call knob: the `agent/request` waterfall proposes a replacement, `prepareCall()` validates it and materializes adapter defaults under the turn signal, and the loop logs the effective value before using the prepared call's registration-bound stream. `callConfigEquals(a, b)` is the field-wise real-change detector; `deepFreeze(value)` is the ownership helper the loop applies to every built request before dispatch (`llm/stream` listeners and adapters read, never rewrite). `markAgentLoopRequest()` gives that exact object process-local loop provenance, and `isAgentLoopRequest()` lets observers distinguish it from independently logged auxiliary calls that may also be frozen and session-associated. `GenerateOptions.purpose` classifies logged auxiliary compaction and session-title calls so adapters can apply purpose-specific transport policy without changing ordinary conversation requests.
+`LlmCallConfig` is the provider, model, optional adapter-owned reasoning effort, and sampling scalars of one conversation's requests (`provider`, `model`, `reasoningEffort`, `temperature`, `maxTokens`, `stop` — each mapping 1:1 onto the same-named `GenerateOptions` field). It is per-conversation state recorded in the session log as part of the request header (see the dsh-session `request/header` events), never a silently-adjustable per-call knob: the `agent/request` waterfall proposes a replacement, `prepareCall()` validates it and materializes adapter defaults under the turn signal, and the loop logs the effective value plus adapter-default provenance before using the prepared call's registration-bound stream. The next proposal omits marked defaults so a changed route resolves its own values; unmarked explicit fields persist. `callConfigEquals(a, b)` is the field-wise real-change detector; `deepFreeze(value)` is the ownership helper the loop applies to every built request before dispatch (`llm/stream` listeners and adapters read, never rewrite). `markAgentLoopRequest()` gives that exact object process-local loop provenance, and `isAgentLoopRequest()` lets observers distinguish it from independently logged auxiliary calls that may also be frozen and session-associated. `GenerateOptions.purpose` classifies logged auxiliary compaction and session-title calls so adapters can apply purpose-specific transport policy without changing ordinary conversation requests.
 
 ### App attribution (`attribution.ts`)
 

packages/llm/llm/README.zh.md

@@ -25,7 +25,7 @@
 
 确切模型元数据是独立的正确性查询，不是 catalog 装饰或全局 LLM 设置。`resolveModelInfo()` 会向拥有精确提供方／模型路由的适配器查询一次；适配器可以描述未列出的动态模型，缺少 `context`、`defaultMaxTokens` 或 `reasoning` 字段会分别保留未知容量、提供方持有的输出默认值或不可用的推理能力。无效的身份、上下文、输出默认值或推理元数据会以 `INVALID_MODEL_INFO`、`INVALID_MODEL_CONTEXT`、`INVALID_MODEL_MAX_TOKENS` 或 `INVALID_MODEL_REASONING` 失败。
 
-`defaultMaxTokens` 是适配器配置的单次请求输出上限，不是模型硬上限。仅当请求省略 `maxTokens` 时，`resolveCallConfig()` 才会填入该值；显式上限优先。推理标识符是由适配器持有的不透明字符串，而非核心枚举：同一次解析只接受与已公布标识符完全一致的值，在存在 `defaultEffort` 时填入它，否则保留提供方默认值。异步模型解析器会接收调用方的 signal，并且必须在取消后迅速结束。`prepareCall()` 还会让精确适配器注册跨越请求头记录和最终分派，因此 HMR（热模块替换）不会将一个适配器的能力结果与另一个适配器的请求混用；复用其一次性句柄或更改调用配置字段会以 `INVALID_PREPARED_CALL` 失败。不支持的显式或配置推理强度会在提供方 I/O 前以 `UNSUPPORTED_REASONING_EFFORT` 失败。
+`defaultMaxTokens` 是适配器配置的单次请求输出上限，不是模型硬上限。仅当请求省略 `maxTokens` 时，`resolveCallConfig()` 才会填入该值；显式上限优先。推理标识符是由适配器持有的不透明字符串，而非核心枚举：同一次解析只接受与已公布标识符完全一致的值，在存在 `defaultEffort` 时填入它，否则保留提供方默认值。异步模型解析器会接收调用方的 signal，并且必须在取消后迅速结束。`prepareCall()` 还会通过 `adapterDefaults` 报告它填入了哪些 `maxTokens` 和 `reasoningEffort` 字段，并让精确适配器注册跨越请求头记录和最终分派，因此 HMR（热模块替换）不会将一个适配器的能力结果与另一个适配器的请求混用；复用其一次性句柄或更改调用配置字段会以 `INVALID_PREPARED_CALL` 失败。不支持的显式或配置推理强度会在提供方 I/O 前以 `UNSUPPORTED_REASONING_EFFORT` 失败。
 
 ### 事件
 
@@ -48,7 +48,7 @@
 
 ### 调用配置（`call-config.ts`）
 
-`LlmCallConfig` 是一个会话中各次请求的提供方、模型、可选的适配器持有推理强度和采样标量（`provider`、`model`、`reasoningEffort`、`temperature`、`maxTokens`、`stop`，每个都与同名 `GenerateOptions` 字段 1:1 映射）。它是作为请求标头一部分记录在会话日志中的每会话状态（见 dsh-session `request/header` 事件），绝不是可静默调整的每次调用旋钮：`agent/request` waterfall 会提议替换，`prepareCall()` 在轮次 signal 控制下校验它并填入适配器默认值，loop 随后记录生效值，再使用已准备调用中与注册绑定的流。`callConfigEquals(a, b)` 是逐字段真实变更检测器；`deepFreeze(value)` 是 loop 在 dispatch 前对每个已构建请求应用的所有权 helper（`llm/stream` listener 与适配器只读，绝不改写）。`markAgentLoopRequest()` 为该精确对象添加进程本地 loop 溯源，`isAgentLoopRequest()` 让观测方可以将其与同样可能冻结并关联会话、但独立记录的辅助调用区分。`GenerateOptions.purpose` 对已记录辅助压缩与会话标题调用分类，让适配器可以应用目的特定传输策略，而不改变普通会话请求。
+`LlmCallConfig` 是一个会话中各次请求的提供方、模型、可选的适配器持有推理强度和采样标量（`provider`、`model`、`reasoningEffort`、`temperature`、`maxTokens`、`stop`，每个都与同名 `GenerateOptions` 字段 1:1 映射）。它是作为请求标头一部分记录在会话日志中的每会话状态（见 dsh-session `request/header` 事件），绝不是可静默调整的每次调用旋钮：`agent/request` waterfall 会提议替换，`prepareCall()` 在轮次 signal 控制下校验它并填入适配器默认值，loop 随后记录生效值及适配器默认值来源，再使用已准备调用中与注册绑定的流。下一次提议会省略带标记的默认值，使变更后的路由解析自身的值；未带标记的显式字段会保留。`callConfigEquals(a, b)` 是逐字段真实变更检测器；`deepFreeze(value)` 是 loop 在 dispatch 前对每个已构建请求应用的所有权 helper（`llm/stream` listener 与适配器只读，绝不改写）。`markAgentLoopRequest()` 为该精确对象添加进程本地 loop 溯源，`isAgentLoopRequest()` 让观测方可以将其与同样可能冻结并关联会话、但独立记录的辅助调用区分。`GenerateOptions.purpose` 对已记录辅助压缩与会话标题调用分类，让适配器可以应用目的特定传输策略，而不改变普通会话请求。
 
 ### 应用归因（`attribution.ts`）
 

packages/llm/llm/src/call-config.ts

@@ -27,6 +27,15 @@ export interface LlmCallConfig {
   stop?: string[]
 }
 
+/**
+ * Effective config fields supplied by exact-model adapter resolution rather
+ * than by the caller's request proposal.
+ */
+export interface LlmCallConfigAdapterDefaults {
+  reasoningEffort?: true
+  maxTokens?: true
+}
+
 /**
  * Field-wise equality over {@link LlmCallConfig} — the comparison a caller
  * runs to decide whether a proposed configuration is a real change (worth a

packages/llm/llm/src/index.ts

@@ -20,7 +20,7 @@ import { resolveRetryPolicy } from './retry-policy.ts'
 import type { ResolvedRetryPolicy } from './retry-policy.ts'
 import type { ProviderRequestId } from './brand.ts'
 import { callConfigEquals, deepFreeze } from './call-config.ts'
-import type { LlmCallConfig } from './call-config.ts'
+import type { LlmCallConfig, LlmCallConfigAdapterDefaults } from './call-config.ts'
 import { HarnessError } from './error.ts'
 import { bindAdapterFailureScope, markLlmAdapterFailure } from './adapter-failure.ts'
 import type { AdapterFailureScope } from './adapter-failure.ts'
@@ -34,7 +34,7 @@ export * from './message.ts'
 export * from './retry-policy.ts'
 export { BlockAssembler } from './assembler.ts'
 export { callConfigEquals, deepFreeze, isAgentLoopRequest, markAgentLoopRequest } from './call-config.ts'
-export type { LlmCallConfig } from './call-config.ts'
+export type { LlmCallConfig, LlmCallConfigAdapterDefaults } from './call-config.ts'
 export { isLlmAdapterFailure, llmFailureOf, llmRetryPolicyOf } from './adapter-failure.ts'
 
 declare module 'cordis' {
@@ -113,6 +113,8 @@ export class LlmError extends HarnessError {
 export interface PreparedLlmCall {
   /** Detached, deep-frozen config with any adapter-owned default materialized. */
   readonly config: LlmCallConfig
+  /** Config fields materialized by the captured adapter rather than proposed by the caller. */
+  readonly adapterDefaults: LlmCallConfigAdapterDefaults
   /**
    * Dispatch this call once through the registration captured during
    * preparation. The request's call-config fields must match {@link config};
@@ -447,12 +449,20 @@ export class LlmService extends Service {
    */
   async prepareCall(config: LlmCallConfig, signal?: AbortSignal): Promise<PreparedLlmCall> {
     const registration = this.registration(config.provider)
-    const resolvedConfig = deepFreeze(structuredClone(
-      await this.resolveCallConfigFor(registration, config, signal),
-    ))
+    const resolved = await this.resolveCallConfigFor(registration, config, signal)
+    const resolvedConfig = deepFreeze(structuredClone(resolved))
+    const adapterDefaults = deepFreeze<LlmCallConfigAdapterDefaults>({
+      ...config.reasoningEffort === undefined && resolved.reasoningEffort !== undefined
+        ? { reasoningEffort: true }
+        : {},
+      ...config.maxTokens === undefined && resolved.maxTokens !== undefined
+        ? { maxTokens: true }
+        : {},
+    })
     let dispatched = false
     return Object.freeze({
       config: resolvedConfig,
+      adapterDefaults,
       stream: (options: GenerateOptions): AsyncIterable<StreamChunk> => {
         if (dispatched) {
           throw new LlmError('a prepared LLM call can only be dispatched once', 'INVALID_PREPARED_CALL')

packages/llm/llm/tests/service.spec.ts

@@ -906,7 +906,12 @@ describe('LlmService', () => {
       { id: 'route', name: 'Route' },
       [],
       {},
-      { model: source },
+      {
+        model: source,
+        providerDefault: {
+          efforts: [{ id: ReasoningEffortId('standard'), name: 'Standard' }],
+        },
+      },
     ))
 
     const resolved = await ctx.llm.resolveModelInfo('route', 'model')
@@ -920,6 +925,8 @@ describe('LlmService', () => {
     })
     const explicit = { provider: 'route', model: 'model', reasoningEffort: ReasoningEffortId('ultra') }
     await expect(ctx.llm.resolveCallConfig(explicit)).resolves.toBe(explicit)
+    const providerDefault = { provider: 'route', model: 'providerDefault' }
+    await expect(ctx.llm.resolveCallConfig(providerDefault)).resolves.toBe(providerDefault)
   })
 
   it('materializes an adapter-owned maxTokens default while preserving an explicit cap', async () => {
@@ -941,8 +948,12 @@ describe('LlmService', () => {
       model: 'model',
       maxTokens: 256_000,
     })
+    const preparedDefault = await ctx.llm.prepareCall({ provider: 'route', model: 'model' })
+    expect(preparedDefault.adapterDefaults).toEqual({ maxTokens: true })
     const explicit = { provider: 'route', model: 'model', maxTokens: 8_192 }
     await expect(ctx.llm.resolveCallConfig(explicit)).resolves.toBe(explicit)
+    const preparedExplicit = await ctx.llm.prepareCall(explicit)
+    expect(preparedExplicit.adapterDefaults).toEqual({})
   })
 
   it.each([0, 1.5, Number.MAX_SAFE_INTEGER + 1])(
@@ -1106,6 +1117,8 @@ describe('LlmService', () => {
     ctx.llm.registerAdapter(['route'], adapter)
     const prepared = await ctx.llm.prepareCall({ provider: 'route', model: 'model' })
     expect(Object.isFrozen(prepared.config)).toBe(true)
+    expect(Object.isFrozen(prepared.adapterDefaults)).toBe(true)
+    expect(prepared.adapterDefaults).toEqual({ reasoningEffort: true })
     const stream = prepared.stream({
       ...prepared.config,
       model: 'other',