fix(web): generate model-backed session titles

.agents/notes/implemented/feature/2026-07-21-log-backed-session-titles.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
-2026-07-21-log-backed-session-titles.md: 6d2aa2049b57d82014f1a85555a1bda9b537bece
-2026-07-21-log-backed-session-titles.zh.md: 7f43832f3a0b6a28227b6a862d78be117c7cb398
+2026-07-21-log-backed-session-titles.md: 183aa6909fecffdaf18c77c2a66fbc38c67c2d2c
+2026-07-21-log-backed-session-titles.zh.md: c6a0c2ce2ad4b2cddec2ada36f655fe55adb143b

.agents/notes/implemented/feature/2026-07-21-log-backed-session-titles.md

@@ -12,7 +12,7 @@ Session identity metadata is immutable, the event log is the replay and fork bou
 
 ## Decision
 
-The [`session-title` capability family](../../../../packages/session-title/README.md) owns title state and generation policy. `@deepseek-ai/dsh-session-title` provides `ctx.sessionTitle`, a deterministic first-message fallback, and a registry for at most one optional asynchronous provider. `@deepseek-ai/dsh-session-title-llm` owns the common auxiliary-model request policy; separate first-message and all-user-messages plugins choose input cadence. The shared agent spine and Web host runtime mount the fallback service with explicit overridable limits; neither composition mounts an asynchronous provider, so a fresh Web session persists a title without adding a model call. Either model provider remains opt-in.
+The [`session-title` capability family](../../../../packages/session-title/README.md) owns title state and generation policy. `@deepseek-ai/dsh-session-title` provides `ctx.sessionTitle`, a deterministic first-message fallback, and a registry for at most one optional asynchronous provider. `@deepseek-ai/dsh-session-title-llm` owns the common auxiliary-model request policy; separate first-message and all-user-messages plugins choose input cadence. The shared agent spine mounts only the fallback service. The Web host mounts that service plus the first-message model provider with explicit overridable limits, so a fresh Web session gains an immediate fallback and then a non-blocking model summary. Other compositions choose either model provider explicitly.
 
 ### Event ownership and folding
 
@@ -32,7 +32,7 @@ The first-message provider schedules once when a fresh session first creates its
 
 `register(provider)` validates one branded stable id, cadence, and generation function, then returns an awaitable effect disposer. A second live registration throws immediately. Provider disposal marks the registration closing, aborts its pending and active work, and waits for every call to settle before removing the registration, so replacement cannot overlap a provider that ignores cancellation. Session disposal aborts its active work. Service teardown prevents queued fallback and provider microtasks from starting, aborts active work, and drains tracked promises before unloading completes. Every session-local generation has a monotonic revision and exact registration identity; acceptance rechecks revision, registration, session liveness, service liveness, and cancellation, so stale output cannot commit.
 
-Model providers require explicit word, CJK-character, input-byte, output-token, and timeout limits. Optional `provider` and `model` overrides are a pair; without them the helper uses the exact route from the logged main request header. Selected messages are framed as JSON under one fixed language-aware instruction. The input limit measures that final user prompt, including wrappers, seq fields, and JSON escaping, before the request is logged or dispatched. Oversized input is rejected rather than truncated because truncation would make the recorded source seqs falsely imply complete use. The fused deadline is checked while consuming each stream chunk and after completion, so a successful result returned after timeout cannot be accepted even when an interceptor or adapter ignores abort.
+Model providers require explicit word, CJK-character, input-byte, output-token, and timeout limits. Optional `provider` and `model` overrides are a pair; without them the helper uses the exact route from the logged main request header. Selected messages are framed as JSON under one fixed language-aware instruction. The dispatched `GenerateOptions` carries `purpose: 'session-title'`; the DeepSeek adapter maps that purpose to thinking-disabled and omits reasoning effort so the bounded output is visible title text, while the main conversation keeps its configured thinking mode. The input limit measures the final user prompt, including wrappers, seq fields, and JSON escaping, before the request is logged or dispatched. Oversized input is rejected rather than truncated because truncation would make the recorded source seqs falsely imply complete use. The fused deadline is checked while consuming each stream chunk and after completion, so a successful result returned after timeout cannot be accepted even when an interceptor or adapter ignores abort.
 
 Automatic provider failures are nonfatal warnings and retain the latest title. Explicit refresh failures reject to the caller. Output must be non-empty text with unique ordered seqs drawn from the fixed request; the service normalizes and byte-limits it before durable acceptance.
 
@@ -50,12 +50,13 @@ A fork inherits seed title events unchanged, like the rest of its source log. Th
 - **Permit multiple registered providers and resolve precedence after completion** — rejected because completion order is not product precedence and would make retries, HMR, and provenance nondeterministic. A deployment that needs a composite policy can register one provider that owns that policy.
 - **Silently truncate oversized auxiliary input** — rejected because the provider result would claim exact source-message provenance while receiving only partial text. Keeping the prior title and warning preserves truthful attribution.
 - **Index titles in `listSessions()` immediately** — rejected because the existing lightweight metadata list would need per-backend derived-index synchronization. Exact `readTitle()` establishes the read contract without precommitting search or indexing policy.
+- **Keep the Web host fallback-only** — rejected because the UI would expose durable titles but never improve them beyond the first-prompt prefix. The first-message provider keeps its latency off the main response path while making model summaries the default Web outcome.
 
 ## Consequences
 
 - Titles survive JSONL and SQLite persistence, replay through ACP, and follow fork inheritance without a separate mutable record.
 - Web title delivery stays incremental and log-backed without a title index or persisted-list scan; cold list rows improve after attach.
-- A fallback appears without an auxiliary call; deployments choose whether better titles justify model cost and whether later prompts should retitle a session.
+- A fallback appears immediately. Each fresh Web session adds one first-message auxiliary call; other compositions choose whether better titles justify model cost and whether later prompts should retitle a session.
 - Auxiliary request records and late accepted titles consume event seqs and may create balanced zero-step turns, so persistence exposes both attempted dispatches and accepted updates even though model history and KV-cache identity do not change.
 - One provider and monotonic per-session revisions make disposal, supersession, and stale-result rejection explicit, at the cost of leaving multi-strategy precedence to a composite provider.
 - Manual rename, deletion, generated-versus-user precedence, search, and list indexing remain outside the capability.

.agents/notes/implemented/feature/2026-07-21-log-backed-session-titles.zh.md

@@ -12,7 +12,7 @@ Status: implemented
 
 ## 决策
 
-[`session-title` 功能包族](../../../../packages/session-title/README.md)负责标题状态和生成策略。`@deepseek-ai/dsh-session-title` 提供 `ctx.sessionTitle`、确定性的首消息回退方案，以及一个至多接受单个可选异步提供方的注册表。`@deepseek-ai/dsh-session-title-llm` 负责通用的辅助模型请求策略；首消息插件和全部用户消息插件分别选择输入调度方式。共享 agent 主干与 Web host 运行时都会挂载回退服务，并显式设置可覆盖的限制；两种组合均不挂载异步提供方，因此新建的 Web 会话无需增加模型调用即可持久化标题。两种模型提供方均仍需按需启用。
+[`session-title` 功能包族](../../../../packages/session-title/README.md)负责标题状态和生成策略。`@deepseek-ai/dsh-session-title` 提供 `ctx.sessionTitle`、确定性的首消息回退方案，以及一个至多接受单个可选异步提供方的注册表。`@deepseek-ai/dsh-session-title-llm` 负责通用的辅助模型请求策略；首消息插件和全部用户消息插件分别选择输入调度方式。共享 agent 主干只挂载回退服务。Web host 会挂载该服务和首消息模型提供方，并显式设置可覆盖的限制，因此新建的 Web 会话会立即获得回退标题，随后在不阻塞主响应的情况下获得模型摘要。其他组合需显式选择任一模型提供方。
 
 ### 事件归属与折叠
 
@@ -32,7 +32,7 @@ Status: implemented
 
 `register(provider)` 会验证一个带品牌类型的稳定 id、执行时机和生成函数，然后返回一个可等待完成的 effect 资源释放函数。第二个活跃注册会立即抛出错误。提供方执行资源释放时，会将注册标记为正在关闭，中止其待执行和活跃工作，并等待所有调用结束后才移除注册，因此替代提供方不会与忽略取消的旧提供方重叠运行。会话资源释放会中止其活跃工作。服务卸载时，会阻止排队中的回退和提供方微任务启动，中止活跃工作，并且卸载完成前会等待所有已跟踪的 promise 结算。每项会话本地生成都有单调递增的修订号和对应的注册身份；接受结果时会重新检查修订号、注册、会话活跃状态、服务活跃状态和取消状态，因此陈旧输出无法提交。
 
-模型提供方必须显式配置单词数、CJK 字符数、输入字节数、输出 token 数和超时限制。可选的 `provider` 和 `model` 覆盖项必须成对提供；两者均未提供时，辅助组件会使用主请求已记录请求头中的准确路由。系统在一条固定且能区分语言的指令下，将选中的消息封装为 JSON。输入字节数按最终形成的用户提示词计算，其中包括包装文本、seq 字段和 JSON 转义；系统会在记录请求或发起调用前完成这项检查。过大输入会被拒绝而不是截断，因为截断会让记录的源消息 seq 错误地表示这些消息已被完整使用。系统在消费每个流分片时以及流完成后都会检查融合后的截止时间，因此即使拦截器或适配器忽略中止信号，超时后返回的成功结果也不会被接受。
+模型提供方必须显式配置单词数、CJK 字符数、输入字节数、输出 token 数和超时限制。可选的 `provider` 和 `model` 覆盖项必须成对提供；两者均未提供时，辅助组件会使用主请求已记录请求头中的准确路由。系统在一条固定且能区分语言的指令下，将选中的消息封装为 JSON。发出的 `GenerateOptions` 携带 `purpose: 'session-title'`；DeepSeek 适配器将该用途映射为禁用思考且省略推理强度设置的请求，使受限输出成为可见的标题文本，而主对话仍沿用已配置的思考模式。输入字节数按最终形成的用户提示词计算，其中包括包装文本、seq 字段和 JSON 转义；系统会在记录请求或发起调用前完成这项检查。过大输入会被拒绝而不是截断，因为截断会让记录的源消息 seq 错误地表示这些消息已被完整使用。系统在消费每个流分片时以及流完成后都会检查融合后的截止时间，因此即使拦截器或适配器忽略中止信号，超时后返回的成功结果也不会被接受。
 
 自动提供方故障只会发出非致命警告，并保留最新标题。显式刷新失败则会向调用方返回拒绝。输出必须是非空文本，并包含来自固定请求、唯一且有序的 seq；服务会在持久接受前对其进行规范化并施加字节限制。
 
@@ -50,12 +50,13 @@ Status: implemented
 - **允许注册多个提供方，并在完成后解析优先级**：不予采纳，因为完成顺序并不等于产品优先级，而且会让重试、HMR 和来源信息变得不确定。需要组合策略的部署可以注册一个自行负责该策略的提供方。
 - **静默截断过大的辅助输入**：不予采纳，因为提供方结果会声明准确的源消息来源信息，实际却只接收了部分文本。保留原有标题并发出警告，可以保持归因真实。
 - **立即在 `listSessions()` 中索引标题**：不予采纳，因为现有的轻量元数据列表将需要逐后端同步派生索引。精确的 `readTitle()` 建立了读取契约，而没有提前锁定搜索或索引策略。
+- **让 Web host 只使用回退标题**：不予采纳，因为 UI 虽会显示持久标题，却始终无法将第一条提示词的前缀改进为更好的标题。首消息提供方在主响应路径之外运行，并让模型摘要成为 Web 的默认结果。
 
 ## 后果
 
 - 标题可以在 JSONL 和 SQLite 持久化中存续，通过 ACP 回放，并遵循 fork 继承语义，而无需单独的可变记录。
 - Web 标题仍以增量方式从日志交付，无需标题索引或扫描持久化列表；冷启动列表项会在会话附加后改用标题。
-- 回退标题无需辅助调用即可出现；部署方可以自行决定更优标题是否值得模型成本，以及后续提示词是否需要重新生成会话标题。
+- 回退标题会立即出现。每个新建的 Web 会话都会增加一次针对首消息的辅助调用；其他组合可以自行决定更优标题是否值得模型成本，以及后续提示词是否需要重新生成会话标题。
 - 辅助请求记录和延迟接受的标题会占用事件 seq，并可能创建平衡的零步骤轮次，因此持久化会同时呈现尝试发起的调用与已接受的更新，尽管模型历史和 KV 缓存标识保持不变。
 - 单个提供方和每会话单调递增的修订号让释放、取代和陈旧结果拒绝行为明确可见，但多策略优先级必须由复合提供方负责。
 - 手动重命名、删除、生成标题与用户标题的优先级、搜索和列表索引不在此功能范围内。

apps/web/tests/smoke-real.e2e.ts

@@ -79,6 +79,7 @@ async function detailsTrack(page: Page): Promise<number> {
 // plugin's client bundle exists and exports apply, the loader fail-louds and
 // the frame never appears.
 const UI_PLUGIN_DIRS = ['connection', 'runtime', 'ui-theme', 'i18n', 'ui-layout', 'ui-sidebar', 'ui-conversation', 'ui-trajectory']
+const ROUND_DONE_MARKER = 'WEB_ROUND_DONE'
 const notReady = UI_PLUGIN_DIRS.filter((dir) => {
   const bundle = join(REPO_ROOT, 'packages/client', dir, 'lib/client.js')
   return !existsSync(bundle) || !readFileSync(bundle, 'utf8').includes('exports.apply')
@@ -175,7 +176,9 @@ describe.skipIf(!process.env.DEEPSEEK_API_KEY || notReady.length > 0)('web smoke
     const input = page.locator('textarea').first()
     await input.waitFor({ timeout: 10_000 })
     await screen(page, '02-empty-state')
-    await input.fill('请简单介绍事件溯源，两句话即可，最后以「介绍完毕」结尾')
+    const prompt = `Please answer this request carefully: explain event sourcing in two sentences, ending with exactly ${ROUND_DONE_MARKER}.`
+    const fallbackTitle = 'Please answer this request carefully:'
+    await input.fill(prompt)
     await input.press('Enter')
     // startSession chain: session mounts, composer moves to the bottom.
     // Regression pin (P0, 585671106): this send used to white-screen the tree
@@ -188,7 +191,14 @@ describe.skipIf(!process.env.DEEPSEEK_API_KEY || notReady.length > 0)('web smoke
       undefined,
       { timeout: 15_000 },
     )
+    await page.waitForFunction(
+      expected => document.title !== `${expected} — DeepSeek Harness`
+        && document.title.endsWith(' — DeepSeek Harness'),
+      fallbackTitle,
+      { timeout: 90_000 },
+    )
     const durableTitle = (await page.title()).replace(/ — DeepSeek Harness$/, '')
+    expect(durableTitle).not.toBe(fallbackTitle)
     const sessionTree = page.getByRole('tree', { name: 'Sessions' })
     const projectRow = sessionTree.getByRole('treeitem').first()
     if (await projectRow.getAttribute('aria-expanded') === 'false') await projectRow.click()
@@ -196,7 +206,7 @@ describe.skipIf(!process.env.DEEPSEEK_API_KEY || notReady.length > 0)('web smoke
       sessionTree.getByText(durableTitle, { exact: true }).waitFor({ timeout: 10_000 }),
       page.getByRole('navigation').getByText(durableTitle, { exact: true }).waitFor({ timeout: 10_000 }),
     ])
-    await page.waitForFunction(() => document.body.innerText.includes('介绍完毕'), undefined, { timeout: 120_000 })
+    await page.waitForFunction(marker => document.body.innerText.includes(marker), ROUND_DONE_MARKER, { timeout: 120_000 })
     await screen(page, '04-round-complete')
   }, 150_000)
 
@@ -271,7 +281,7 @@ describe.skipIf(!process.env.DEEPSEEK_API_KEY || notReady.length > 0)('web smoke
     onTestFailed(() => saveFailureShot(page, 'w5-reload'))
     await page.reload({ waitUntil: 'load' })
     await page.waitForSelector('[class*="frame"]', { timeout: 30_000 })
-    await page.waitForFunction(() => document.body.innerText.includes('介绍完毕'), undefined, { timeout: 30_000 })
+    await page.waitForFunction(marker => document.body.innerText.includes(marker), ROUND_DONE_MARKER, { timeout: 30_000 })
     await screen(page, '12-reload-recovery')
   })
 

docs/core-data-structures/core.md

@@ -238,10 +238,10 @@ interface GenerateOptions {
   sessionId?: Branded<'SessionId'>
   /**
    * Provider-neutral classification for an auxiliary model call. Adapters may
-   * map the purpose to model-hidden transport metadata. Ordinary conversation
-   * requests leave it unset.
+   * map the purpose to model-hidden transport metadata or purpose-specific
+   * generation policy. Ordinary conversation requests leave it unset.
    */
-  purpose?: 'compaction'
+  purpose?: 'compaction' | 'session-title'
 }
 ```
 

packages/cordis/tool-cordis/src/api-catalog.ts

@@ -1409,7 +1409,7 @@ export const TYPE_API: readonly TypeApiEntry[] = [
   },
   {
     name: 'GenerateOptions',
-    declaration: 'export interface GenerateOptions {\n    provider: string;\n    model: string;\n    messages: Message[];\n    system?: string;\n    tools?: ToolSchema[];\n    temperature?: number;\n    maxTokens?: number;\n    stop?: string[];\n    signal?: AbortSignal;\n    sessionId?: Branded<\'SessionId\'>;\n    purpose?: \'compaction\';\n}',
+    declaration: 'export interface GenerateOptions {\n    provider: string;\n    model: string;\n    messages: Message[];\n    system?: string;\n    tools?: ToolSchema[];\n    temperature?: number;\n    maxTokens?: number;\n    stop?: string[];\n    signal?: AbortSignal;\n    sessionId?: Branded<\'SessionId\'>;\n    purpose?: \'compaction\' | \'session-title\';\n}',
   },
   {
     name: 'GenericCallView',

packages/host/runtime/README.md

@@ -1,6 +1,6 @@
 # @deepseek-ai/dsh-host-runtime
 
-Host runtime assembly for `dsh`: `bootHost` composes the core plugin spine (LLM service + DeepSeek adapter, sessions with JSONL persistence and deterministic fallback titles, system prompt, tools, agents, agent loop, local bash), `createApiProxy` implements the [`dsh-host-apiproxy`](../apiproxy/README.md) contract over that composition, and `startHost` is the one-step shell seam returning `{ api, handler, defaults, ctx, dispose }`.
+Host runtime assembly for `dsh`: `bootHost` composes the core plugin spine (LLM service + DeepSeek adapter, sessions with JSONL persistence, immediate fallback titles and first-message model summaries, system prompt, tools, agents, agent loop, local bash), `createApiProxy` implements the [`dsh-host-apiproxy`](../apiproxy/README.md) contract over that composition, and `startHost` is the one-step shell seam returning `{ api, handler, defaults, ctx, dispose }`.
 
 Which plugins mount and with what defaults is decided only here — shells must not `ctx.plugin` to alter the assembly. `RunningHost.ctx` is a formal seam with exactly two sanctioned uses: mounting protocol front-door plugins (e.g. a future `dsh acp`) and headless session-event subscription; consuming clients must not bypass `api` through it.
 
@@ -12,7 +12,8 @@ Which plugins mount and with what defaults is decided only here — shells must
 | `provider` | `'deepseek'` | Default provider route injected as agentOptions on create/resume and reported by `host.describe`. |
 | `model` | `'deepseek-v4-flash'` | Default model id, same single source as `provider`. |
 | `cwd` | `process.cwd()` | Default project directory for a session whose create request omits `cwd`. |
-| `sessionTitle` | 5 words / 40 fallback bytes / 80 accepted bytes | Deterministic fallback-title limits. The host mounts no asynchronous title provider, so title creation adds no model call. |
+| `sessionTitle` | 5 words / 40 fallback bytes / 80 accepted bytes | Deterministic fallback and accepted-title limits. |
+| `sessionTitleLlm` | 5 words / 10 CJK chars / 4,096 input bytes / 64 output tokens / 60 s | First-message model-title policy. An omitted route inherits the logged main-request provider and model. |
 
 ## ApiProxy implementation notes
 
@@ -20,11 +21,11 @@ Unary methods take the narrow `RpcRequest<P>` and echo `request.rpcId`; a prompt
 
 ## Model Experience
 
-Indirectly, through the model-facing plugins bootHost mounts and the provider/model defaults injected into created and resumed agents.
+Indirectly, through the non-blocking first-message title request owned by [`dsh-session-title-llm`](../../session-title/session-title-llm/README.md) and the other model-facing plugins `bootHost` mounts.
 
 #### KV Cache effect
 
-No direct invalidation; the mounted model-facing plugins own their request-prefix changes.
+No main-request invalidation; the auxiliary title request has its own cache behavior and the conversation prefix remains unchanged.
 
 ## Known Limitations and Deferred Work
 

packages/host/runtime/package.json

@@ -50,6 +50,7 @@
     "@deepseek-ai/dsh-session-persistence": "workspace:^",
     "@deepseek-ai/dsh-session-persistence-jsonl": "workspace:^",
     "@deepseek-ai/dsh-session-title": "workspace:^",
+    "@deepseek-ai/dsh-session-title-first-message-llm": "workspace:^",
     "@deepseek-ai/dsh-skill": "workspace:^",
     "@deepseek-ai/dsh-skill-local": "workspace:^",
     "@deepseek-ai/dsh-spill-local": "workspace:^",

packages/host/runtime/src/boot.ts

@@ -9,6 +9,8 @@ import Timer from '@cordisjs/plugin-timer'
 import LlmService from '@deepseek-ai/dsh-llm'
 import SessionStore from '@deepseek-ai/dsh-session'
 import SessionTitleService, { type Config as SessionTitleConfig } from '@deepseek-ai/dsh-session-title'
+import * as SessionTitleFirstMessageLlm from '@deepseek-ai/dsh-session-title-first-message-llm'
+import type { Config as SessionTitleLlmConfig } from '@deepseek-ai/dsh-session-title-first-message-llm'
 import SystemPrompt from '@deepseek-ai/dsh-system-prompt'
 import ToolRegistry from '@deepseek-ai/dsh-tools'
 import AgentRegistry from '@deepseek-ai/dsh-agent'
@@ -46,6 +48,15 @@ const DEFAULT_SESSION_TITLE_CONFIG: SessionTitleConfig = {
   maxTitleBytes: 80,
 }
 
+/** Default first-message model-title policy for sessions created through the host. */
+const DEFAULT_SESSION_TITLE_LLM_CONFIG: SessionTitleLlmConfig = {
+  targetWords: 5,
+  targetCjkCharacters: 10,
+  maxInputBytes: 4_096,
+  maxOutputTokens: 64,
+  timeoutMs: 60_000,
+}
+
 /** Options for bootHost — the assembly-layer composition knobs. */
 export interface BootHostOptions {
   /** Root directory for JSONL session persistence. */
@@ -54,8 +65,10 @@ export interface BootHostOptions {
   provider?: string
   /** Default model id (defaults to 'deepseek-v4-flash', matching the demos). */
   model?: string
-  /** Deterministic fallback-title limits; no asynchronous title provider is mounted by the host. */
+  /** Deterministic fallback-title limits. */
   sessionTitle?: SessionTitleConfig
+  /** First-message model-title policy; omitted provider/model inherit the session's logged main-request route. */
+  sessionTitleLlm?: SessionTitleLlmConfig
   /**
    * Default project directory for sessions created without an explicit cwd
    * (defaults to the host process working directory). A session's cwd is its
@@ -100,6 +113,7 @@ export async function bootHost(options: BootHostOptions): Promise<HostHandle> {
   await ctx.plugin(LlmService)
   await ctx.plugin(SessionStore)
   await ctx.plugin(SessionTitleService, options.sessionTitle ?? DEFAULT_SESSION_TITLE_CONFIG)
+  await ctx.plugin(SessionTitleFirstMessageLlm, options.sessionTitleLlm ?? DEFAULT_SESSION_TITLE_LLM_CONFIG)
   await ctx.plugin(SystemPrompt, { persona: '' })
   await ctx.plugin(ToolRegistry)
   await ctx.plugin(AgentRegistry)

packages/host/runtime/tests/host-runtime.spec.ts

@@ -9,6 +9,7 @@ import type { GenerateOptions, StreamChunk } from '@deepseek-ai/dsh-llm'
 import { LlmAdapter } from '@deepseek-ai/dsh-llm'
 import type { SessionId } from '@deepseek-ai/dsh-session'
 import type { Config as SessionTitleConfig } from '@deepseek-ai/dsh-session-title'
+import type { Config as SessionTitleLlmConfig } from '@deepseek-ai/dsh-session-title-first-message-llm'
 import type { HostFrame, MuxFrame } from '@deepseek-ai/dsh-host-apiproxy/api'
 import type { RpcRequest, RpcResponse } from '@deepseek-ai/dsh-host-apiproxy/api/rpc'
 import { RpcId } from '@deepseek-ai/dsh-host-apiproxy/api/rpc'
@@ -21,6 +22,10 @@ class ScriptedAdapter extends LlmAdapter {
   }
 
   async * stream(options: GenerateOptions): AsyncIterable<StreamChunk> {
+    if ((options.tools?.length ?? 0) === 0) {
+      yield * textResponse('Durable append-only session titles')
+      return
+    }
     const entry = this.script.shift()
     if (!entry) throw new Error('ScriptedAdapter: script exhausted')
     if (entry === 'hang') {
@@ -96,6 +101,7 @@ afterEach(async () => {
 async function boot(
   script: (StreamChunk[] | 'hang')[] = [],
   sessionTitle?: SessionTitleConfig,
+  sessionTitleLlm?: SessionTitleLlmConfig,
 ): Promise<RunningHost> {
   host = await startHost({
     boot: {
@@ -103,6 +109,7 @@ async function boot(
       provider: 'scripted',
       model: 'test-model',
       ...(sessionTitle === undefined ? {} : { sessionTitle }),
+      ...(sessionTitleLlm === undefined ? {} : { sessionTitleLlm }),
     },
   })
   host.ctx.llm.registerAdapter(['scripted'], new ScriptedAdapter(script))
@@ -158,6 +165,63 @@ describe('sessions.create / list', () => {
 })
 
 describe('sessions.prompt / cancel', () => {
+  it.each([
+    { name: 'host default', config: undefined, target: '5 words', maxTokens: 64 },
+    {
+      name: 'configured policy',
+      config: {
+        targetWords: 3,
+        targetCjkCharacters: 8,
+        maxInputBytes: 2_048,
+        maxOutputTokens: 24,
+        timeoutMs: 2_000,
+      },
+      target: '3 words',
+      maxTokens: 24,
+    },
+  ] satisfies {
+    name: string
+    config: SessionTitleLlmConfig | undefined
+    target: string
+    maxTokens: number
+  }[])('replaces the fallback with a model-backed first-message title using the $name', async ({ config, target, maxTokens }) => {
+    const modelTitle = 'Durable append-only session titles'
+    const running = await boot([textResponse('pong')], undefined, config)
+    const { api, ctx } = running
+    const { sessionId } = expectOk(await api.sessions.create(request({})))
+    const agent = ctx.agents.get(sessionId) as Agent
+    const idle = waitForIdle(ctx, agent)
+    expectOk(await api.sessions.prompt(request({
+      sessionId,
+      mode: 'queue' as const,
+      content: [{ type: 'text' as const, text: 'Explain why append-only logs make session titles durable.' }],
+    })))
+    await idle
+
+    await vi.waitFor(() => {
+      expect(agent.session.events.filter(event => event.type === 'session/title').map(event => event.data))
+        .toEqual([
+          {
+            title: 'Explain why append-only logs make',
+            messageSeqs: [1],
+            source: { kind: 'fallback' },
+          },
+          {
+            title: modelTitle,
+            messageSeqs: [1],
+            source: {
+              kind: 'provider',
+              provider: 'session-title-first-message-llm',
+              model: { provider: 'scripted', model: 'test-model' },
+            },
+          },
+        ])
+    })
+    const titleRequest = agent.session.events.find(event => event.type === 'session/title-llm-request')
+    expect(titleRequest?.data.system).toContain(target)
+    expect(titleRequest?.data.maxTokens).toBe(maxTokens)
+  })
+
   it.each([
     { name: 'host default', config: undefined, expected: 'Show the Web UI durable' },
     {

packages/host/runtime/tsconfig.json

@@ -26,6 +26,9 @@
     {
       "path": "../../session-title/session-title"
     },
+    {
+      "path": "../../session-title/session-title-first-message-llm"
+    },
     {
       "path": "../../core/system-prompt"
     },

packages/llm/llm-deepseek/README.md

@@ -32,7 +32,7 @@ The plugin registers the single provider route `deepseek`. A request selects it
 
 `reasoningEffort` is **omitted by default** — when unset, the `reasoning_effort` wire field is not sent and the server applies its own default for the model. The only accepted values are `high` and `max` (DeepSeek's official effort levels). It is meaningful only with thinking enabled (the provider default).
 
-`thinking`/`reasoningEffort` are adapter-level request defaults serialized as the official top-level `thinking: {type}` / `reasoning_effort` wire fields. They live in adapter config (not `GenerateOptions`) to keep the core vocabulary provider-neutral.
+`thinking`/`reasoningEffort` are adapter-level request defaults serialized as the official top-level `thinking: {type}` / `reasoning_effort` wire fields. They live in adapter config (not `GenerateOptions`) to keep the core vocabulary provider-neutral. A request with `GenerateOptions.purpose: 'session-title'` forces thinking disabled and omits `reasoning_effort`, reserving its bounded output for visible title text without changing conversation or compaction defaults.
 
 `streamIdleTimeoutMs` bounds each outstanding provider read, including the initial `fetch`, without counting time the consumer spends between chunks. One stable abort signal reaches the request and body reader for the whole call; expiry stops the transport and throws `LlmError('TIMEOUT')`, while an earlier caller abort throws `LlmError('ABORTED')`. The adapter makes exactly one provider request per `stream()` call; agent-level retry is a separate plugin policy.
 

packages/llm/llm-deepseek/src/serialize.ts

@@ -118,14 +118,18 @@ export function serializeRequest(options: GenerateOptions, defaults: RequestDefa
       parameters: tool.parameters,
     },
   }))
+  // A short title budget must produce visible text; conversation and
+  // compaction calls continue to inherit the adapter's thinking defaults.
+  const thinking = options.purpose === 'session-title' ? 'disabled' : defaults.thinking
+  const reasoningEffort = options.purpose === 'session-title' ? undefined : defaults.reasoningEffort
 
   return {
     model: options.model,
     messages,
     stream: true,
     stream_options: { include_usage: true },
-    ...defaults.thinking !== undefined ? { thinking: { type: defaults.thinking } } : {},
-    ...defaults.reasoningEffort !== undefined ? { reasoning_effort: defaults.reasoningEffort } : {},
+    ...thinking !== undefined ? { thinking: { type: thinking } } : {},
+    ...reasoningEffort !== undefined ? { reasoning_effort: reasoningEffort } : {},
     ...tools !== undefined && tools.length > 0 ? { tools } : {},
     ...options.temperature !== undefined ? { temperature: options.temperature } : {},
     ...options.maxTokens !== undefined ? { max_tokens: options.maxTokens } : {},

packages/llm/llm-deepseek/tests/serialize.spec.ts

@@ -180,6 +180,15 @@ describe('serializeRequest', () => {
     expect(wire.reasoning_effort).toBe('max')
   })
 
+  it('disables thinking for session-title requests without changing adapter defaults', () => {
+    const wire = serializeRequest(
+      request({ messages: history, purpose: 'session-title' }),
+      { thinking: 'enabled', reasoningEffort: 'max' },
+    )
+    expect(wire.thinking).toEqual({ type: 'disabled' })
+    expect(wire.reasoning_effort).toBeUndefined()
+  })
+
   it('omits thinking fields when unset (provider default applies)', () => {
     const wire = serializeRequest(request({ messages: history }))
     expect(wire.thinking).toBeUndefined()

packages/llm/llm/README.md

@@ -39,7 +39,7 @@ Streaming is a raw chunk protocol (`block-start`, `text-delta`, `reasoning-delta
 
 ### Call configuration (`call-config.ts`)
 
-`LlmCallConfig` is the provider + model + sampling scalars of one conversation's requests (`provider`, `model`, `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 and the loop logs a real change. `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.
+`LlmCallConfig` is the provider + model + sampling scalars of one conversation's requests (`provider`, `model`, `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 and the loop logs a real change. `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/src/types.ts

@@ -228,8 +228,8 @@ export interface GenerateOptions {
   sessionId?: Branded<'SessionId'>
   /**
    * Provider-neutral classification for an auxiliary model call. Adapters may
-   * map the purpose to model-hidden transport metadata. Ordinary conversation
-   * requests leave it unset.
+   * map the purpose to model-hidden transport metadata or purpose-specific
+   * generation policy. Ordinary conversation requests leave it unset.
    */
-  purpose?: 'compaction'
+  purpose?: 'compaction' | 'session-title'
 }

packages/session-title/session-title-llm/README.md

@@ -8,7 +8,7 @@ This package is a library, not a Cordis plugin. The provider plugins call `regis
 
 `provider` and `model` overrides are optional but must be supplied together as non-empty strings. Without that pair, the helper uses the exact provider/model route captured from the current session's logged `request/header`; an explicit refresh before any route exists therefore needs overrides. The helper measures the final JSON-framed user prompt, including seq fields, wrappers, and JSON escaping, against `maxInputBytes` before logging or dispatch instead of truncating it. Timeout and caller cancellation are rechecked while consuming the stream and after it completes, so a late successful result cannot be accepted even if an interceptor or adapter ignores abort. Malformed or empty output, tool calls, and non-stop finish reasons also reject; the session-title service decides whether that rejection is an automatic warning or an explicit caller failure.
 
-After route and input validation, the helper appends a log-only `session/title-llm-request` event before model dispatch. It contains the title-provider id, exact source seqs, route, system prompt, message list, and output-token cap used by the call. The append shares the title capability's per-session settlement queue, so a superseding request cannot collide with an earlier fallback, request record, or accepted-title flush. The dispatched envelope is deep-frozen to keep interceptors aligned with that record but deliberately lacks dsh-agent-loop's process-local request identity, so loop-only reconstruction observers do not compare it with the conversation header. A later model failure leaves that request record intact; validation failures that never become dispatchable requests do not create one. The event stays outside derived model history.
+After route and input validation, the helper appends a log-only `session/title-llm-request` event before model dispatch. It contains the title-provider id, exact source seqs, route, system prompt, message list, and output-token cap used by the call. The append shares the title capability's per-session settlement queue, so a superseding request cannot collide with an earlier fallback, request record, or accepted-title flush. The dispatched envelope is deep-frozen, carries `purpose: 'session-title'`, and deliberately lacks dsh-agent-loop's process-local request identity. Interceptors stay aligned with the record while loop-only reconstruction observers do not compare it with the conversation header. The DeepSeek adapter maps that purpose to thinking-disabled so the small output budget is reserved for visible title text; other adapters own their purpose-specific behavior. A later model failure leaves the request record intact; validation failures that never become dispatchable requests do not create one. The event stays outside derived model history.
 
 ## Configuration
 
@@ -33,7 +33,7 @@ The title model receives a fixed system instruction to return one concise unador
 
 #### Token effect
 
-The auxiliary request consumes tokens according to selected input size and `maxOutputTokens`. It is separate from the main agent request and does not add title text or framing to agent history.
+The auxiliary request consumes tokens according to selected input size and `maxOutputTokens`. It is separate from the main agent request and does not add title text or framing to agent history. DeepSeek title calls disable thinking; the main conversation retains its configured thinking mode.
 
 #### KV Cache effect
 

packages/session-title/session-title-llm/src/index.ts

@@ -261,6 +261,7 @@ export async function generateSessionTitleWithLlm(
     system,
     maxTokens: config.maxOutputTokens,
     sessionId: request.session.id,
+    purpose: 'session-title',
     signal: callDeadline.signal,
   })
   await appendSessionTitleOutOfBand(ctx, request.session, 'session/title-llm-request', {

packages/session-title/session-title-llm/tests/llm.spec.ts

@@ -156,6 +156,7 @@ describe('generateSessionTitleWithLlm', () => {
       model: 'current-model',
       maxTokens: 32,
       sessionId: providerRequest.session.id,
+      purpose: 'session-title',
     })
     expect(options.system).toContain('5 words')
     expect(options.system).toContain('10 CJK characters')

pnpm-lock.yaml

@@ -2035,6 +2035,9 @@ importers:
       '@deepseek-ai/dsh-session-title':
         specifier: workspace:^
         version: link:../../session-title/session-title
+      '@deepseek-ai/dsh-session-title-first-message-llm':
+        specifier: workspace:^
+        version: link:../../session-title/session-title-first-message-llm
       '@deepseek-ai/dsh-skill':
         specifier: workspace:^
         version: link:../../skill/skill