page-agent-sdk 3.22.0 → 3.23.0

package/README.md

@@ -8,7 +8,7 @@
 
 [![npm](https://img.shields.io/npm/v/page-agent-sdk.svg)](https://www.npmjs.com/package/page-agent-sdk)
 [![license](https://img.shields.io/badge/license-ISC-blue.svg)](https://github.com/whyymj/page-agent-sdk/blob/master/LICENSE)
-[![tests](https://img.shields.io/badge/self%20tests-2319%20asserts-brightgreen.svg)](#self-tests)
+[![tests](https://img.shields.io/badge/self%20tests-2377%20asserts-brightgreen.svg)](#self-tests)
 
 ---
 
@@ -139,6 +139,7 @@ CDN zero-config: `<script src="https://unpkg.com/page-agent-sdk"></script>` →
 | 👁 DOM read (2.20+) | Read rendered DOM structure (depth-cutoff + attr whitelist); verify modifications took effect — distinct from `eval_script` (structured + read-only) | `capabilities.domInspect` |
 | 📊 Context inspector | Snapshot actual-LLM-message composition (total / occupancy / category ratio); DebugDrawer `📊 上下文` tab + `inspectContext()`; zero LLM cost, default on | `capabilities.contextInspector` |
 | 🤖 Agent-driven compression (2.33+) | `capabilities.agentCompression` (opt-in) lets the summary LLM decide per-trigger compression strategy via an `inspect_context` tool loop (keepRounds / windowRatio / summary mode / recall / preserve); `shouldTriggerCompression` gate avoids per-message LLM cost; decide failure/timeout degrades to static; `decisionTimeoutMs` / `decisionMaxTokens` configurable | `capabilities.agentCompression` + `summaryLlm` |
+| 🎯 Cross-session user preference memory | `capabilities.preferences` (**opt-in, default off** — auto-writing the user's browser is behavior-sensitive): the agent captures durable user preferences from conversation — strong signal (explicit commands like "Remember: …", zero LLM) / medium signal (pattern-word prefilter + small-LLM extraction; the core test is **durable taste vs this-round task instruction**) / behavioral inference **not captured** (better to miss than to learn wrong — one false preference would ride along every future session); preferences persist independently (preferenceStore, IndexedDB, same shape as storage/skillStorage; same topic **later statement overrides earlier**, FIFO ≤20); injected as a pin segment into the system prompt each round (survives sessions and compression); manage wrongly-learned entries via `sdk.getPreferences()/removePreference(id)/clearPreferences()`, plus a read-only DebugDrawer "User preferences" section | `capabilities: { preferences: true }` + optional `preferenceStorage` |
 | ⚡ host actions (2.20+) | Register save/publish/preview etc; SDK auto-generates named tools, agent triggers page ops directly (no `trigger_action` indirection) | `actions` |
 | 🧩 schema tiered disclosure (2.20+) | Large schema → systemPrompt injects top-level overview only (no constraints/no recursion); deep constraints via `schema_data` on demand; small schema unaffected (full) | `schemaHint` |
 | 📌 cross-compress working memory (2.20+) | Pin recent read/query paths + hashes across compression; no re-fetch, correct optimistic-lock hash | `capabilities.workingMemory` |
@@ -146,7 +147,7 @@ CDN zero-config: `<script src="https://unpkg.com/page-agent-sdk"></script>` →
 | 📐 context resilience (2.30+) | Hard floor `contextWindow ≥200K` (rejects <200K models like legacy `deepseek`/`gpt-4o`/`glm-4.5` at startup); three gates (compress/trim/offload) thresholds follow the live window after `setLlm`; reactive retry on `context_length_exceeded` (aggressive trim → single retry, never fails raw); vfs large-result refs protected from LRU eviction + OOM 1.5× fallback; system-prompt budget (25% window, drops non-pinned segments, keeps base/mission/workingMemory) | built-in |
 | 🎯 focus auto-switch (2.31+) | AI auto-judges task scope → `set_focus` (local task) / `clear_focus` (global/done); focus persists across refresh/session-switch (restore validates path via `getSchemaAtPath`, drops if invalid); spawned subagents inherit parent focus (three-layer convergence; parent unfocused → child no focus middleware, zero regression) | `capabilities.focus` + `toolMode:'advanced'` |
 | 🔒 precise-value protection (2.32+) | `data.resources: [{path, mode}]` protects exact-value fields: `freeze` (read-only, value hidden via `⟦frozen:path⟧` placeholder, FROZEN_FIELD on write) / `verbatim` (preserved verbatim, `⟦res:handle⟧`, modify via `resource_update` else VERBATIM_MISMATCH); write-side enforcement across commitSetToBind/applyPatches/eval + resource tools (`resource_get/update/list/delete`, advanced) + cross-compression pin | `data.resources` + `capabilities.vfs` |
-| 🌍 UI customization & i18n (3.17+–3.22+) | Dialog UI is fully customizable without forking: `dialog.icons` (per-icon override; plain text or sanitized HTML fragment) + built-in `dialog.theme: 'dark'`, and the top-level `i18n` group (3.22+): `locale: 'en-US'` switches the built-in message pack (chat surface + Debug drawer + Skill panel + code preview; `formatTime`/autoTitle follow, **default systemPrompt switches to English** so agent replies match the UI language), `messages` per-key overrides (e.g. `statusDone: '<b style="color:#10b981">Done ✓</b>'` — rich-text render spots accept inline HTML fragments sanitized via a text allowlist) — switch language and tweak individual strings in one group; `DialogMessages` (~219 keys) + `MESSAGES_ZH_CN`/`MESSAGES_EN_US`/`resolveDialogMessages` exported for custom UIs | `dialog.{icons,theme}` + `i18n.{locale,messages}` |
+| 🌍 UI customization & i18n (3.17+–3.22+) | Dialog UI is fully customizable without forking: `dialog.icons` (per-icon override; plain text or sanitized HTML fragment) + built-in `dialog.theme: 'dark'`, and the top-level `i18n` group (3.22+): `locale: 'en-US'` switches the built-in message pack (chat surface + Debug drawer + Skill panel + code preview; `formatTime`/autoTitle follow, **default systemPrompt switches to English** so agent replies match the UI language), `messages` per-key overrides (e.g. `statusDone: '<b style="color:#10b981">Done ✓</b>'` — rich-text render spots accept inline HTML fragments sanitized via a text allowlist) — switch language and tweak individual strings in one group; `DialogMessages` (~226 keys) + `MESSAGES_ZH_CN`/`MESSAGES_EN_US`/`resolveDialogMessages` exported for custom UIs | `dialog.{icons,theme}` + `i18n.{locale,messages}` |
 
 Capabilities default on (`verify`/`approval`/`checkpoint` default off; **proactive `humanConfirm` default on** — AI asks when uncertain/multi-plan instead of guessing). Turn off unneeded ones via `capabilities` to save tokens.
 
@@ -213,7 +214,7 @@ ChatDialog, MessageContent, CodePreview, SkillPanel, DebugDrawer, useChat
 | | `augmentSystem` | `(ctx:{state,data?}) => string \| undefined` | Dynamic system prompt injection hook: called each turn, returns a string injected as a segment based on runtime state/data; return undefined to skip; callback errors degrade to skip (no crash). `ctx.data` is taken from liveData() each turn (auto-syncs after setData), enabling dynamic component descriptions / partial schema hints. Not set = current behavior |
 | **Page data** | `data` | `{schema,bind,description?}` | Single main object: declare zod schema (validation + field descriptions auto-injected into prompt) + bind (reactive/plain object, tools read/write directly, no `window`) + description |
 | | `tools` / `skills` / `memory` | `Tool[]` / `SkillSpec[]` / `string` | Custom tools / skills / AGENTS.md-style directives |
-| **Capability toggles** | `capabilities` | `{planning?,dataOps?,fetch?,skills?,vfs?,summarization?,memory?,subagent?,verify?,focus?}` | Default all on (`verify` default off; `focus` = context focus for refining one component, default on); `false` to turn off |
+| **Capability toggles** | `capabilities` | `{planning?,dataOps?,fetch?,skills?,vfs?,summarization?,memory?,subagent?,verify?,focus?,preferences?}` | Default all on (`verify`/`preferences` default off, opt-in; `focus` = context focus for refining one component, default on; `preferences` = cross-session preference memory); `false` to turn off |
 | | `permissions` | `PermissionRule[]` | Scope whitelist (first-match-wins, default off) |
 | | `humanConfirm` | `boolean` · default `true` | Proactive inquiry (AI asks when uncertain/multi-plan) |
 | | `approval` | `{tools?,confirm?,timeoutMs?,humanConfirmTool?}` · default off | Passive confirm whitelist (pre-write allow/deny) |
@@ -249,7 +250,7 @@ ChatDialog, MessageContent, CodePreview, SkillPanel, DebugDrawer, useChat
 | `inputRows` | `number` · default `2` | Input box rows (visible height); `1` = single row; `2` = 2-row initial height, auto-expands up to max-height:100px; `>2` = taller initial height |
 | `onClose` | `() => void` | Drawer mode close callback (default `hide`; pass to override and sync external mount state) |
 | `theme` | `'light' \| 'dark'` · default `'dark'` | Built-in theme (dark = Ark design palette); fully customizable via `--cs-*` on an ancestor |
-| `i18n` | `I18nOptions` | **Top-level i18n group (3.22+; replaces `dialog.locale`/`dialog.messages`)**: `locale` switches the built-in message pack — chat surface + Debug drawer + Skill panel + CodePreview; `formatTime` (12h/24h), autoTitle, and the **default systemPrompt** follow (`en-US` → English `DEFAULT_SYSTEM_PROMPT_EN` with a "Respond in English" anchor, so agent replies match the UI language; a custom `systemPrompt` is untouched, but the auto-appended `reliableWriteRules` segment switches to English). `messages` = per-key overrides (priority over the locale pack, e.g. `statusDone: '<b style="color:#10b981">Done ✓</b>'` — values may be inline HTML fragments on rich-text render spots, sanitized via a text allowlist). Full key list (~219 keys) in `DialogMessages` |
+| `i18n` | `I18nOptions` | **Top-level i18n group (3.22+; replaces `dialog.locale`/`dialog.messages`)**: `locale` switches the built-in message pack — chat surface + Debug drawer + Skill panel + CodePreview; `formatTime` (12h/24h), autoTitle, and the **default systemPrompt** follow (`en-US` → English `DEFAULT_SYSTEM_PROMPT_EN` with a "Respond in English" anchor, so agent replies match the UI language; a custom `systemPrompt` is untouched, but the auto-appended `reliableWriteRules` segment switches to English). `messages` = per-key overrides (priority over the locale pack, e.g. `statusDone: '<b style="color:#10b981">Done ✓</b>'` — values may be inline HTML fragments on rich-text render spots, sanitized via a text allowlist). Full key list (~226 keys) in `DialogMessages` |
 | `icons` | `Partial<DialogIcons>` | **Icon customization**: partial override of default emojis (`header` 🤖 / `subagent` 🤖 / `subagentProgress` 🧬 / `empty` 💬 / `focus` 🎯 / `queued` 📋 / `queuedEdit` ✏️ / `recommend` 💡 / `conflict` ⚠️; `assistantAvatar`/`userAvatar` default to built-in SVG, pass a text glyph to replace). Values: plain text (emoji/char) **or an HTML fragment** (starting with `<`, e.g. inline `<svg>`/`<img>` — sanitized via a DOMPurify icon allowlist; event attributes/dangerous protocols stripped); empty string hides the icon; unset keys keep defaults |
 
 ### Extension points
@@ -532,8 +533,8 @@ function switchTo(i: number) {
 ## Self-tests
 
 ```bash
-npm test            # 2319 assertions (tsx, source-level; no LLM dependency)
-npm run test:e2e    # 718 integration assertions (node, built dist; covers APIs/options/modules/simple&complex scenes: default systemPrompt(capability overview) / dynamic register + inspect sync / inspect(tools/middleware/subagent/verify/mcp/todos/lastCompression/checkpoints reflect config) / custom tools/middleware/skills/memory injection / runtime dynamic reconfiguration(setTools/addTool/removeTool/setLlm/setMemory/setSubagents reflect) / switchSession(on/off) / shareContext on/off sharing/independent / storage backends + object config / presets(3) / checkpoint / exports complete(39+ fns/components) / util fns usable(isQuotaError/estimateTokens/jpEval/searchJson) / source=builtin / mount boundary / hook multi-listener / llm config / hide/show / error scenes)
+npm test            # 2377 assertions (tsx, source-level; no LLM dependency)
+npm run test:e2e    # 737 integration assertions (node, built dist; covers APIs/options/modules/simple&complex scenes: default systemPrompt(capability overview) / dynamic register + inspect sync / inspect(tools/middleware/subagent/verify/mcp/todos/lastCompression/checkpoints reflect config) / custom tools/middleware/skills/memory injection / runtime dynamic reconfiguration(setTools/addTool/removeTool/setLlm/setMemory/setSubagents reflect) / switchSession(on/off) / shareContext on/off sharing/independent / storage backends + object config / presets(3) / checkpoint / exports complete(39+ fns/components) / util fns usable(isQuotaError/estimateTokens/jpEval/searchJson) / source=builtin / mount boundary / hook multi-listener / llm config / hide/show / error scenes)
 ```
 
 ## Local npm package test

package/README.zh-CN.md

@@ -8,7 +8,7 @@
 
 [![npm](https://img.shields.io/npm/v/page-agent-sdk.svg)](https://www.npmjs.com/package/page-agent-sdk)
 [![license](https://img.shields.io/badge/license-ISC-blue.svg)](https://github.com/whyymj/page-agent-sdk/blob/master/LICENSE)
-[![tests](https://img.shields.io/badge/self%20tests-2319%20asserts-brightgreen.svg)](#自测)
+[![tests](https://img.shields.io/badge/self%20tests-2377%20asserts-brightgreen.svg)](#自测)
 
 ---
 
@@ -140,7 +140,8 @@ CDN 零配置：`<script src="https://unpkg.com/page-agent-sdk"></script>` → `
 | 📐 上下文健壮性 (2.30+) | 硬地板 `contextWindow ≥200K`(启动拒绝 <200K 模型如老款 `deepseek`/`gpt-4o`/`glm-4.5`);三道闸(压缩/trim/offload)阈值在 `setLlm` 后跟随实时窗口;遇 `context_length_exceeded` 反应性重试(激进 trim → 重试一次,不裸失败);vfs 大结果引用受保护免 LRU 淘汰 + OOM 1.5× 兜底;系统段预算(25% 窗口,丢弃非 pin 段保 base/mission/workingMemory) | 内置 |
 | 🎯 focus 自动切换 (2.31+) | AI 自动判断任务范围 → `set_focus`(局部任务)/ `clear_focus`(全局/完成);focus 跨刷新/切会话持久化(restore 经 `getSchemaAtPath` 校验 path,失效丢弃);子 agent 继承主焦点(三层收敛;主未聚焦 → 子无 focus 中间件,零回归) | `capabilities.focus` + `toolMode:'advanced'` |
 | 🔒 精确值保护 (2.32+) | `data.resources: [{path, mode}]` 保护需精确保存字段:`freeze`(只读,精确值经 `⟦frozen:path⟧` 占位符不入消息流,写撞 FROZEN_FIELD)/ `verbatim`(原样保留,`⟦res:handle⟧`,改值经 `resource_update` 否则 VERBATIM_MISMATCH);写侧强制覆盖 commitSetToBind/applyPatches/eval + 资源工具(`resource_get/update/list/delete`,advanced)+ 跨压缩 pin | `data.resources` + `capabilities.vfs` |
-| 🌍 UI 定制与国际化 (3.17+~3.22+) | 对话框 UI 免 fork 全定制:`dialog.icons` 逐图标覆盖(纯文本或净化后 HTML 片段)+ 内置深色主题 `dialog.theme:'dark'` + **顶层 `i18n` 配置组(3.22+)**:`locale:'en-US'` 切内置文案包(聊天面 + Debug 抽屉 + Skill 面板 + 代码预览;`formatTime`/autoTitle 跟随,**默认 systemPrompt 切英文** → agent 回复语言与 UI 一致)、`messages` 键级覆盖(如 `statusDone: '<b style="color:#10b981">Done ✓</b>'` —— 富文本渲染位支持行内 HTML 片段,文案白名单净化)——换语言与改个别文案一个配置组;`DialogMessages`(~219 键)+ `MESSAGES_ZH_CN`/`MESSAGES_EN_US`/`resolveDialogMessages` 导出供自建 UI 复用 | `dialog.{icons,theme}` + `i18n.{locale,messages}` |
+| 🌍 UI 定制与国际化 (3.17+~3.22+) | 对话框 UI 免 fork 全定制:`dialog.icons` 逐图标覆盖(纯文本或净化后 HTML 片段)+ 内置深色主题 `dialog.theme:'dark'` + **顶层 `i18n` 配置组(3.22+)**:`locale:'en-US'` 切内置文案包(聊天面 + Debug 抽屉 + Skill 面板 + 代码预览;`formatTime`/autoTitle 跟随,**默认 systemPrompt 切英文** → agent 回复语言与 UI 一致)、`messages` 键级覆盖(如 `statusDone: '<b style="color:#10b981">Done ✓</b>'` —— 富文本渲染位支持行内 HTML 片段,文案白名单净化)——换语言与改个别文案一个配置组;`DialogMessages`(~226 键)+ `MESSAGES_ZH_CN`/`MESSAGES_EN_US`/`resolveDialogMessages` 导出供自建 UI 复用 | `dialog.{icons,theme}` + `i18n.{locale,messages}` |
+| 🎯 跨会话用户偏好记忆 | `capabilities.preferences`(**opt-in 默认关**,自动写用户浏览器属行为敏感项):agent 从对话中捕获用户持久偏好 —— 强信号(「记住:…」显式命令,零 LLM)/ 中信号(模式词初筛 + 小 LLM 提炼,核心判定「持久口味 vs 本轮任务指令」)/ 行为推断**不捕获**(宁漏勿误:学错一条假偏好,之后每个会话都带着跑);偏好独立持久化(preferenceStore,IndexedDB,与 storage/skillStorage 同构;同 topic **后说覆盖前说**,FIFO ≤20);每轮经 pin 段注入 system prompt(跨会话/跨压缩生效);`sdk.getPreferences()/removePreference(id)/clearPreferences()` 管理学错条目,DebugDrawer「用户偏好」只读小节可查 | `capabilities: { preferences: true }` + 可选 `preferenceStorage` |
 
 能力默认开（`verify`/`approval`/`checkpoint` 默认关；**主动征询 `humanConfirm` 默认开**——AI 遇不确定/多方案主动问你、不猜测），可经 `capabilities` 关掉无用的省 token。
 
@@ -207,7 +208,7 @@ ChatDialog, MessageContent, CodePreview, SkillPanel, DebugDrawer, useChat
 | | `augmentSystem` | `(ctx:{state,data?}) => string \| undefined` | 动态 system prompt 注入钩子:每轮调,按运行时 state/data 返回字符串作为一段注入;返回 undefined 跳过;回调抛错降级跳过(不崩)。`ctx.data` 每轮从 liveData() 取最新(setData 后自动同步),可据此动态算当前组件说明 / 部分 schema 描述。不配 = 现状行为 |
 | **页面数据** | `data` | `{schema,bind,description?}` | 单主对象:声明 zod schema(校验 + 字段描述自动注入提示词)+ bind(reactive/普通对象,工具直接读写,不挂 window)+ description |
 | | `tools` / `skills` / `memory` | `Tool[]` / `SkillSpec[]` / `string` | 自定义工具 / 技能 / AGENTS.md 风格持久指令 |
-| **能力开关** | `capabilities` | `{planning?,missionAnchor?,dataOps?,fetch?,skills?,vfs?,summarization?,memory?,workingMemory?,subagent?,verify?,domInspect?,focus?}` | 核心默认开（`verify`/`domInspect` 默认关,opt-in;`focus` 上下文聚焦·指定组件精修,默认开)；`false` 关掉省 token |
+| **能力开关** | `capabilities` | `{planning?,missionAnchor?,dataOps?,fetch?,skills?,vfs?,summarization?,memory?,workingMemory?,subagent?,verify?,domInspect?,focus?,preferences?}` | 核心默认开（`verify`/`domInspect`/`preferences` 默认关,opt-in;`focus` 上下文聚焦·指定组件精修,默认开;`preferences` 跨会话偏好记忆)；`false` 关掉省 token |
 | | `actions` | `Record<string,{description,run,params?}>` | **(2.18+) 宿主动作**：注册 save_draft/publish 等页面操作 → SDK 自动生成命名 tool 供 agent 触发 |
 | | `schemaHint` | `{maxKeys?,maxChars?}` · 默认 `{15,4000}` | **(2.18+) 大 schema 分层披露阈值**：超则 systemPrompt 只注入顶层概览（不带约束/不递归）,深层约束按需 `schema_data` 查;小 schema 无感（全量） |
 | | `permissions` | `PermissionRule[]` | scope 白名单（first-match-wins，默认不启用） |
@@ -245,7 +246,7 @@ ChatDialog, MessageContent, CodePreview, SkillPanel, DebugDrawer, useChat
 | `inputRows` | `number` · 默认 `2` | 输入框行数(可见高度);`1` = 单行;`2` = 2 行初始高度,自动扩展至 max-height:100px;`>2` = 更高初始高度 |
 | `onClose` | `() => void` | 抽屉模式关闭回调(默认 `hide`;传此选项覆盖默认,便于同步外部挂载状态) |
 | `theme` | `'light' \| 'dark'` · 默认 `'dark'` | 内置主题(dark = 方舟设计稿深色紫调);亦可祖先覆盖 `--cs-*` 完全自定义 |
-| `i18n` | `I18nOptions` | **顶层国际化配置组(3.22+;原 `dialog.locale`/`dialog.messages` 两键合并至此)**:`locale` 切换内置文案包 —— 聊天面 + Debug 抽屉 + Skill 面板 + 代码预览;`formatTime`(12h/24h)、autoTitle 与**默认 systemPrompt** 跟随(`en-US` → 英文版 `DEFAULT_SYSTEM_PROMPT_EN` 含 "Respond in English" 语言锚,agent 回复语言与 UI 一致;自定义 `systemPrompt` 不受影响,但其自动追加的 `reliableWriteRules` 段切英文)。`messages` = 键级覆盖(优先于 locale 包,如 `statusDone: '<b style="color:#10b981">完成</b>'` —— 富文本渲染位的值支持行内 HTML 片段,文案白名单净化);完整键清单(~219 键)见 `DialogMessages` |
+| `i18n` | `I18nOptions` | **顶层国际化配置组(3.22+;原 `dialog.locale`/`dialog.messages` 两键合并至此)**:`locale` 切换内置文案包 —— 聊天面 + Debug 抽屉 + Skill 面板 + 代码预览;`formatTime`(12h/24h)、autoTitle 与**默认 systemPrompt** 跟随(`en-US` → 英文版 `DEFAULT_SYSTEM_PROMPT_EN` 含 "Respond in English" 语言锚,agent 回复语言与 UI 一致;自定义 `systemPrompt` 不受影响,但其自动追加的 `reliableWriteRules` 段切英文)。`messages` = 键级覆盖(优先于 locale 包,如 `statusDone: '<b style="color:#10b981">完成</b>'` —— 富文本渲染位的值支持行内 HTML 片段,文案白名单净化);完整键清单(~226 键)见 `DialogMessages` |
 | `icons` | `Partial<DialogIcons>` | **图标自定义**:局部覆盖默认 emoji(`header` 🤖 / `subagent` 🤖 / `subagentProgress` 🧬 / `empty` 💬 / `focus` 🎯 / `queued` 📋 / `queuedEdit` ✏️ / `recommend` 💡 / `conflict` ⚠️;`assistantAvatar`/`userAvatar` 缺省 = 内置 SVG,传文本字形替换)。值为纯文本(emoji/字符)或 **HTML 片段**(以 `<` 开头,如内联 `<svg>`/`<img>`,经 DOMPurify 图标白名单净化,事件属性/危险协议剥除);空串 = 隐藏该图标;未传键用默认 |
 
 ### 扩展点
@@ -477,8 +478,8 @@ function switchTo(i: number) {
 ## 自测
 
 ```bash
-npm test            # 2319 项断言（tsx 源码级，不依赖 LLM）
-npm run test:e2e    # 718 项集成断言（node 跑构建产物 dist；覆盖各 API/配置项/功能模块/简单与复杂场景：默认 systemPrompt(含能力概述) / 动态注册与 inspect 同步 / inspect(tools/middleware/subagent/verify/mcp/todos/lastCompression/checkpoints 反映配置) / 自定义 tools/middleware/skills/memory 注入 / 运行时动态重配置(setTools/addTool/removeTool/setLlm/setMemory/setSubagents 反映) / switchSession(开/未开) / shareContext 开/关共享独立 / storage 后端+对象配置 / presets 三预设 / checkpoint / 导出项完整(39+ 函数/组件) / 工具函数可用(isQuotaError/estimateTokens/jpEval/searchJson) / source=builtin / mount 边界 / hook 多监听器 / llm 配置 / 错误场景）
+npm test            # 2377 项断言（tsx 源码级，不依赖 LLM）
+npm run test:e2e    # 737 项集成断言（node 跑构建产物 dist；覆盖各 API/配置项/功能模块/简单与复杂场景：默认 systemPrompt(含能力概述) / 动态注册与 inspect 同步 / inspect(tools/middleware/subagent/verify/mcp/todos/lastCompression/checkpoints 反映配置) / 自定义 tools/middleware/skills/memory 注入 / 运行时动态重配置(setTools/addTool/removeTool/setLlm/setMemory/setSubagents 反映) / switchSession(开/未开) / shareContext 开/关共享独立 / storage 后端+对象配置 / presets 三预设 / checkpoint / 导出项完整(39+ 函数/组件) / 工具函数可用(isQuotaError/estimateTokens/jpEval/searchJson) / source=builtin / mount 边界 / hook 多监听器 / llm 配置 / 错误场景）
 ```
 
 ## 本地 npm 包测试