fix: show expandable project and conversation usage totals

Group usage into project or conversation totals with collapsed Runtime/model children. Keep model grouping flat and preserve expanded rows when usage refreshes. Include aggregation and interaction tests plus matching documentation.

Validation: 25 focused tests, typecheck, and lint passed. Full suite timed out after 200 seconds with four failures reported in ipc.test.ts and ApplicationMenu.test.tsx.

Release note: Project and conversation usage now show totals first, with model details available by expanding each row.

FEATURES.md

@@ -696,7 +696,9 @@ records are listed separately and do not introduce another feature status.
   without a persistent legacy-cache truncation warning; previously lost
   history is not restored. Token usage is grouped by Runtime and
   model and normalizes the different OpenAI-compatible and Anthropic Messages
-  cache-reporting semantics when showing cache hit rate. Activity is grouped
+  cache-reporting semantics when showing cache hit rate. Project and conversation
+  usage groups show collapsed totals by default and expand to Runtime/model child
+  rows; model grouping remains flat. Activity is grouped
   by conversation and collapsed by default so long histories do not fill the
   page.
 - [x] **Compact conversation tool records**: Expand individual tool records

FEATURES.zh-CN.md

@@ -107,7 +107,7 @@
 
 ### 工作管理、长期协作与工作流
 
-- [x] **任务、活动与成果**：集中管理任务状态、审计活动和独立成果文件；普通聊天回复只保留在会话中，不再自动复制到成果栏，已有重复聊天 Markdown 从成果列表隐藏但不物理删除。运行记录改由 Main SQLite 保存，不再受旧 Renderer 的 500 条、单详情 4,000 字符或约 2 MB 上限影响，页面按批继续显示；不再常驻旧缓存截断警告，但不会恢复此前已丢失的历史。Token 用量按 Runtime 与模型归类，并针对 OpenAI 兼容与 Anthropic Messages 的不同上报口径归一化展示缓存命中率；活动按会话分组并默认收起，避免长历史占满页面。
+- [x] **任务、活动与成果**：集中管理任务状态、审计活动和独立成果文件；普通聊天回复只保留在会话中，不再自动复制到成果栏，已有重复聊天 Markdown 从成果列表隐藏但不物理删除。运行记录改由 Main SQLite 保存，不再受旧 Renderer 的 500 条、单详情 4,000 字符或约 2 MB 上限影响，页面按批继续显示；不再常驻旧缓存截断警告，但不会恢复此前已丢失的历史。Token 用量按 Runtime 与模型归类，并针对 OpenAI 兼容与 Anthropic Messages 的不同上报口径归一化展示缓存命中率；用量按项目或会话分组时默认收起并显示汇总，展开后显示 Runtime / 模型子行，按模型分组时保持平铺；活动按会话分组并默认收起，避免长历史占满页面。
 - [x] **紧凑会话工具记录**：逐项展开工具记录即可查看和复制结果、错误及输入参数；会话与子任务进度使用简洁行，减少重复 Runtime 摘要。OpenCode、Continue、DeepSeek Harness 和直连模型可显示已有文件路径、命令或搜索摘要，并在工具完成和会话重载后保留。
 - [x] **Task 与定制任务体验**：每个产品级 Task 只关联一条 Conversation，一条 Conversation 可承载多个 Task；左侧会话列表通过行首展开按钮显示带共享状态点的 Task 子项，父会话行不重复任务标签，UI 只展示到 Task，不暴露 Job/Run 层级。新建定制任务可关联当前或新 Conversation，默认 Execute 并沿用 Runtime、工具和审批边界；重复触发复用同一 Task，文本结果回写 Conversation，独立文件和图片保留为成果。普通消息与到期 Scheduled Task 共用 Conversation 级持久发送队列，同一会话一次只执行一项；当前回复期间仍可继续发送，队列按顺序续跑，并允许删除或“立即中断并插入”。Task Center 继续作为完整索引，不建设独立 Automation Center。当前计划触发支持单次、每日和每周；高级时区、Cron、事件触发与重试治理仍按 PRD 逐步实现。详见 [Task Center PRD](./docs/features/task-and-job/task-center-prd.md) 和 [Scheduled Task PRD](./docs/features/task-and-job/scheduled-task-prd.md)。
 - [x] **记忆与智能心跳**：当前提供周期回顾、建议记忆、洞察、后续任务和可审计运行轨迹。

UI-DESIGN.md

@@ -738,7 +738,7 @@ GoodBuddy 是可调整窗口大小的桌面应用。响应式设计优先保证
 - 默认视图按“项目 → 任务或会话 → 活动详情”组织，项目范围持续可见，任务或会话详情可以折叠。
 - 任务或会话的综合状态以最近一次顶层请求对应的最终 Agent 结果为准；最终结果尚未产生时使用该请求的当前状态。中间工具或子专家的失败、取消和中断保留在活动详情中，但不得覆盖最终成功状态。
 - 活动时间线按项目分组、按任务或会话建立横向轨道，所有轨道共享同一执行顺序并按事件时间排列，以带身份名称的节点表达用户、主 Agent、子专家、工具、审批、状态和并行关系。节点内使用 `U / G / S / T / A` 拉丁字母简称，节点下显示完整身份或名称；选择节点后显示所属会话、不可变范围快照和完整详情。
-- 用量统计与活动记录分离，支持按项目、会话和模型切换统计维度，宽表格在独立容器内横向滚动。
+- 用量统计与活动记录分离，支持按项目、会话和模型切换统计维度；按项目或会话分组时默认收起并显示汇总，展开后显示 Runtime / 模型子行，按模型分组时保持平铺。宽表格在独立容器内横向滚动。
 - 活动状态筛选使用 `SegmentedControl`，不与页面页签混合。清空历史遵循破坏性操作政策。
 
 ### 13.7 魔法笔记

src/renderer/src/ActivityPanel.test.tsx

@@ -777,7 +777,7 @@ describe('ActivityPanel', () => {
     expect(within(stats).getByText('32%')).toBeInTheDocument()
 
     const projectRow = screen.getByRole('row', {
-      name: '项目甲直连模型 · gpt-5 100 20 10 40 40% 120'
+      name: '项目甲 100 20 10 40 40% 120'
     })
     expect(projectRow).toBeInTheDocument()
     expect(
@@ -810,6 +810,59 @@ describe('ActivityPanel', () => {
     ).toBeInTheDocument()
   })
 
+  it('expands model children under project and conversation totals and keeps them current', () => {
+    const usage = makeTokenUsage()
+    usage.records.push({
+      ...usage.records[0]!,
+      requestId: 'request-other-model',
+      model: 'gpt-other'
+    })
+    const props = {
+      onClear: vi.fn(),
+      onOpenConversation: vi.fn(),
+      records: [],
+      tokenUsage: usage
+    }
+    const { rerender } = render(<ActivityPanel {...props} />)
+    fireEvent.click(screen.getByRole('tab', { name: '用量统计' }))
+    const project = screen.getByRole('button', { name: '项目甲' })
+    expect(project).toHaveAttribute('aria-expanded', 'false')
+    expect(screen.queryByText('直连模型 · gpt-5')).not.toBeInTheDocument()
+    expect(screen.getByRole('row', {
+      name: '项目甲 200 40 20 80 40% 240'
+    })).toBeInTheDocument()
+    fireEvent.click(project)
+    expect(project).toHaveAttribute('aria-expanded', 'true')
+    expect(screen.getByText('直连模型 · gpt-5')).toBeInTheDocument()
+    expect(screen.getByText('直连模型 · gpt-other')).toBeInTheDocument()
+    const updated = {
+      ...usage,
+      records: [...usage.records, { ...usage.records[0]!, requestId: 'new' }]
+    }
+    rerender(<ActivityPanel {...props} tokenUsage={updated} />)
+    expect(screen.getByRole('row', {
+      name: '项目甲 300 60 30 120 40% 360'
+    })).toBeInTheDocument()
+    expect(screen.getByRole('row', {
+      name: '直连模型 · gpt-5 200 40 20 80 40% 240'
+    })).toBeInTheDocument()
+    fireEvent.click(project)
+    expect(screen.queryByText('直连模型 · gpt-5')).not.toBeInTheDocument()
+
+    fireEvent.click(screen.getByRole('button', { name: '按会话' }))
+    const conversation = screen.getByRole('button', { name: '会话甲' })
+    expect(conversation).toHaveAttribute('aria-expanded', 'false')
+    fireEvent.click(conversation)
+    expect(screen.getByText('直连模型 · gpt-other')).toBeInTheDocument()
+    fireEvent.click(conversation)
+    expect(screen.queryByText('直连模型 · gpt-other')).not.toBeInTheDocument()
+
+    fireEvent.click(screen.getByRole('button', { name: '按模型' }))
+    const table = screen.getByRole('table')
+    expect(within(table).queryByRole('button')).not.toBeInTheDocument()
+    expect(within(table).getByText('直连模型 · gpt-5')).toBeInTheDocument()
+  })
+
   it('shows Runtime names instead of internal provider identifiers', () => {
     const usage = makeTokenUsage()
     usage.records.push({

src/renderer/src/ActivityPanel.tsx

@@ -1,4 +1,4 @@
-import { Activity, Trash2 } from 'lucide-react'
+import { Activity, ChevronRight, Trash2 } from 'lucide-react'
 import { useMemo, useState } from 'react'
 import { useTranslation } from 'react-i18next'
 import { InlineHelp } from './InlineHelp'
@@ -253,6 +253,9 @@ export function ActivityPanel({
   const [filter, setFilter] = useState<ActivityFilter>('all')
   const [tokenGroup, setTokenGroup] =
     useState<TokenUsageGroup>('project')
+  const [expandedTokenRows, setExpandedTokenRows] = useState<Set<string>>(
+    () => new Set()
+  )
   const [confirmingClear, setConfirmingClear] = useState(false)
   const [selectedTimelineRecordId, setSelectedTimelineRecordId] =
     useState<string>()
@@ -529,7 +532,7 @@ export function ActivityPanel({
   ]
   const localizeTokenRow = (
     row: (typeof tokenRows)[number]
-  ): { label: string; detail?: string } => {
+  ): { label: string } => {
     const modelLabel =
       row.model || t('tokenUsage.fallbacks.unknownModel')
     const runtimeLabel =
@@ -544,19 +547,15 @@ export function ActivityPanel({
               : row.runtime || t('tokenUsage.fallbacks.unknownRuntime')
     const label =
       tokenGroup === 'project' &&
-      row.key.startsWith('project:unassigned:')
+      row.key === 'project:unassigned'
         ? t('tokenUsage.fallbacks.unassignedProject')
         : tokenGroup === 'conversation' &&
-            row.key.startsWith('conversation:deleted:')
+            row.key === 'conversation:deleted'
           ? t('tokenUsage.fallbacks.deletedConversation')
-          : tokenGroup === 'model'
+          : !row.children
             ? `${runtimeLabel} · ${modelLabel}`
             : row.label
-    const detail =
+    return { label }
-      tokenGroup === 'model'
-        ? undefined
-        : `${runtimeLabel} · ${modelLabel}`
-    return { label, detail }
   }
   const renderRecordCard = (
     record: ActivityRecord,
@@ -1086,14 +1085,44 @@ export function ActivityPanel({
                       </td>
                     </tr>
                   ) : (
-                    tokenRows.map((row) => {
+                    tokenRows.flatMap((parent) => [
+                      { row: parent, isChild: false },
+                      ...(expandedTokenRows.has(parent.key)
+                        ? (parent.children ?? []).map((row) => ({
+                            row: { ...row, key: `${parent.key}:${row.key}` },
+                            isChild: true
+                          }))
+                        : [])
+                    ]).map(({ row, isChild }) => {
                       const localizedRow = localizeTokenRow(row)
                       return (
-                        <tr key={row.key}>
+                        <tr
+                          key={row.key}
+                          className={isChild ? 'token-usage__model-row' : undefined}
+                        >
                           <th scope="row">
-                            <span>{localizedRow.label}</span>
+                            {row.children ? (
-                            {localizedRow.detail && (
+                              <button
-                              <small>{localizedRow.detail}</small>
+                                type="button"
+                                className="token-usage__expand"
+                                aria-expanded={expandedTokenRows.has(row.key)}
+                                onClick={() =>
+                                  setExpandedTokenRows((current) => {
+                                    const next = new Set(current)
+                                    if (next.has(row.key)) {
+                                      next.delete(row.key)
+                                    } else {
+                                      next.add(row.key)
+                                    }
+                                    return next
+                                  })
+                                }
+                              >
+                                <ChevronRight aria-hidden="true" size={14} />
+                                <span>{localizedRow.label}</span>
+                              </button>
+                            ) : (
+                              <span>{localizedRow.label}</span>
                             )}
                           </th>
                           <td>

src/renderer/src/styles.css

@@ -14786,6 +14786,43 @@ details.settings-section > :not(summary) + :not(summary) {
   text-align: left;
 }
 
+.token-usage__expand {
+  display: flex;
+  align-items: center;
+  gap: var(--space-2);
+  width: 100%;
+  padding: var(--space-1) 0;
+  border: 0;
+  background: transparent;
+  color: inherit;
+  font: inherit;
+  text-align: left;
+  cursor: pointer;
+}
+
+.token-usage__expand:hover {
+  color: var(--accent);
+}
+
+.token-usage__expand:focus-visible {
+  outline: 2px solid var(--accent);
+  outline-offset: 2px;
+}
+
+.token-usage__expand svg {
+  flex-shrink: 0;
+}
+
+.token-usage__expand[aria-expanded='true'] svg {
+  transform: rotate(90deg);
+}
+
+.token-usage .token-usage__model-row th:first-child {
+  padding-left: var(--space-6);
+  color: var(--text-secondary);
+  font-weight: 400;
+}
+
 .token-usage tbody th span,
 .token-usage tbody th small {
   display: block;

src/renderer/src/token-usage.test.ts

@@ -67,10 +67,11 @@ describe('token usage aggregation', () => {
     })
     expect(groupTokenUsage(usage, 'project')).toEqual([
       {
-        key: 'project:project-1:runtime:model:model:gpt-5',
+        key: 'project:project-1',
         label: '项目一',
-        model: 'gpt-5',
+        model: '',
-        runtime: 'model',
+        runtime: '',
+        children: groupTokenUsage(usage, 'model'),
         inputTokens: 112,
         outputTokens: 23,
         cacheReadTokens: 40,
@@ -143,7 +144,7 @@ describe('token usage aggregation', () => {
     expect(groupTokenUsage(usage, 'model')[0]?.label).toBe('')
   })
 
-  it('keeps project and conversation totals separated by model', () => {
+  it('aggregates project and conversation totals with model children', () => {
     const usage = makeTokenUsage()
     usage.records.push({
       ...usage.records[0]!,
@@ -154,8 +155,22 @@ describe('token usage aggregation', () => {
       output: 2
     })
 
-    expect(groupTokenUsage(usage, 'project')).toHaveLength(2)
+    const projects = groupTokenUsage(usage, 'project')
-    expect(groupTokenUsage(usage, 'conversation')).toHaveLength(3)
+    expect(projects).toHaveLength(1)
+    expect(projects[0]).toMatchObject({
+      inputTokens: 119,
+      outputTokens: 25,
+      cacheInputTokens: 169,
+      cacheHitRate: 80 / 169,
+      totalTokens: 144
+    })
+    expect(projects[0]?.children).toEqual(groupTokenUsage(usage, 'model'))
+    const conversations = groupTokenUsage(usage, 'conversation')
+    expect(conversations).toHaveLength(2)
+    expect(conversations[0]?.children).toHaveLength(2)
+    expect(conversations[0]?.totalTokens).toBe(129)
+    expect(conversations[1]?.children).toHaveLength(1)
+    expect(conversations[1]?.totalTokens).toBe(15)
   })
 
   it('groups identical model names by Runtime instead of provider', () => {

src/renderer/src/token-usage.ts

@@ -17,6 +17,7 @@ export type TokenUsageGroupRow = TokenUsageTotals & {
   label: string
   model: string
   runtime: string
+  children?: TokenUsageGroupRow[]
 }
 
 type TokenUsageRecord = TokenUsageSummary['records'][number]
@@ -80,10 +81,10 @@ function groupIdentity(
       ? `project:${record.projectId}`
       : 'project:unassigned'
     return {
-      key: `${projectKey}:${runtimeModelKey}`,
+      key: projectKey,
       label: record.projectName?.trim() || '',
-      model,
+      model: '',
-      runtime
+      runtime: ''
     }
   }
 
@@ -92,10 +93,10 @@ function groupIdentity(
       ? `conversation:${record.conversationId}`
       : 'conversation:deleted'
     return {
-      key: `${conversationKey}:${runtimeModelKey}`,
+      key: conversationKey,
       label: record.conversationTitle?.trim() || '',
-      model,
+      model: '',
-      runtime
+      runtime: ''
     }
   }
 
@@ -118,9 +119,15 @@ export function groupTokenUsage(
   group: TokenUsageGroup
 ): TokenUsageGroupRow[] {
   const rows = new Map<string, TokenUsageGroupRow>()
+  const groupedRecords = new Map<string, TokenUsageRecord[]>()
 
   for (const record of tokenUsage.records) {
     const identity = groupIdentity(record, group)
+    if (group !== 'model') {
+      const records = groupedRecords.get(identity.key) ?? []
+      records.push(record)
+      groupedRecords.set(identity.key, records)
+    }
     const usage = usageNumbers(record, record.provider)
     const existing = rows.get(identity.key)
 
@@ -151,5 +158,15 @@ export function groupTokenUsage(
     })
   }
 
-  return [...rows.values()]
+  return [...rows.values()].map((row) =>
+    group === 'model'
+      ? row
+      : {
+          ...row,
+          children: groupTokenUsage(
+            { ...tokenUsage, records: groupedRecords.get(row.key)! },
+            'model'
+          )
+        }
+  )
 }