Fixture/flick (#1280)

* fix: 终端内容溢出 viewport 时的重影 bug 主屏幕模式下 frame 持续溢出 viewport 时，cursor-restore LF 把内容滚入 scrollback 导致相对光标追踪漂移，可见区 diff 落到错误行产生重影（重复 banner / 错位）。扩展 log-update overflow 分支为无条件 fullReset（含 \x1b[3J 清 scrollback），并将主屏 self-healing 清屏从 ERASE_SCREEN (CSI 2 J) 换成 ERASE_DOWN (CSI J)，避免 xterm.js / VSCode 集成终端的 scrollback 边界副作用。 Co-Authored-By: glm-5.2 <zai-org@claude-code-best.win> * chore: 删除 3 个孤立诊断脚本 - scripts/verify-autofix-pr.ts: 一次性 autofix-pr 验证脚本，全仓零引用 - scripts/smoke-test-commands.ts: 开发期冒烟测试脚本，无任何 import - scripts/probe-subscription-endpoints.ts: 手动 endpoint 探针，无引用均不在 package.json scripts、build.ts、vite.config.ts、CI workflows 中。 Co-Authored-By: glm-5.2 <zai-org@claude-code-best.win> * chore: 移除 self-hosted-runner stub 及其 cli.tsx fast-path - 删除 src/self-hosted-runner/main.ts（自动生成的 Promise.resolve() stub） - 同步移除 src/entrypoints/cli.tsx 中 feature('SELF_HOSTED_RUNNER') 守卫的 fast-path 分支 - 该 flag 不在 build.ts DEFAULT_BUILD_FEATURES 也不在 dev 默认列表，所有默认配置下整段为构建期死代码删除 stub 单独会留下未解析的动态 import，必须协同拆除。 Co-Authored-By: glm-5.2 <zai-org@claude-code-best.win> * chore: 删除 agentSdkTypes 中三个 not-implemented stub 移除 watchScheduledTasks、buildMissedTaskNotification、connectRemoteControl 三个 stub 函数（函数体仅 throw new Error('not implemented')），以及仅被这些 stub 引用的孤儿类型（ScheduledTasksHandle、ConnectRemoteControlOptions、RemoteControlHandle、InboundPrompt 等）。全仓零外部引用。buildMissedTaskNotification 在 src/utils/cronScheduler.ts 有真实可用实现，未受影响。 Co-Authored-By: glm-5.2 <zai-org@claude-code-best.win> * chore: 移除 Cursor.ts 中未引用的 kill ring 访问器 - 删除 getKillRingItem、getKillRingSize、clearKillRing、canYankPop（全仓零引用的独立 export） - 移除 VIM_WORD_CHAR_REGEX 的 export 关键字（仍由 isVimWordChar 内部使用，保留常量本体） kill ring 特性本身仍活跃（getLastKill/pushToKillRing/yankPop 在 useSearchInput/useTextInput 使用），仅这几个孤儿 helper 未接入。 Co-Authored-By: glm-5.2 <zai-org@claude-code-best.win> * chore: 移除 insights.ts 中未引用的导出 - 删除 deduplicateSessionBranches（全仓零调用，含 JSDoc） - 删除 buildExportData（全仓零调用，原 S3 上传路径实际用 HTML 而非 JSON） - InsightsExport 仅移除 export 关键字（保留类型本体，仍作为内部返回类型） Co-Authored-By: glm-5.2 <zai-org@claude-code-best.win> * chore: 移除 autonomyCommandSpec.ts 中未引用的导出 - 删除 AUTONOMY_CLI（CLI 子命令描述对象，零引用；handler 仅用 AUTONOMY_USAGE） - 删除 AUTONOMY_COMMAND_DESCRIPTION（值已在 main.tsx:5181 内联） - ParsedAutonomyCommand 仅移除 export 关键字（保留类型作为 parseAutonomyArgs 返回类型） Co-Authored-By: glm-5.2 <zai-org@claude-code-best.win> * chore: 移除 binaryCheck/claudeAiLimits/codeIndexing 中未引用的导出 - binaryCheck.ts: 删除 clearBinaryCache（零调用，binaryCache 仍由 isBinaryInstalled 使用） - claudeAiLimits.ts: 删除 RATE_LIMIT_DISPLAY_NAMES 常量 + getRateLimitDisplayName（互为唯一消费者） - codeIndexing.ts: 删除 detectCodeIndexingFromMcpTool（同胞 detectCodeIndexingFromCommand/McpServerName 仍活跃） Co-Authored-By: glm-5.2 <zai-org@claude-code-best.win> * chore: 移除多处仅内部使用的 export 关键字下列符号均仅在本文件内被引用，export 关键字冗余；保留符号本体不动： - internalLogging.ts: getContainerId（line 88 内部调用） - api/errors.ts: isMediaSizeError（line 151 内部调用） - api/withRetry.ts: parseMaxTokensContextOverflowError（line 389/724 内部调用） - statsCache.ts: STATS_CACHE_VERSION（7 处内部使用） - startupProfiler.ts: logStartupPerf（line 128 内部调用） - bashCommandHelpers.ts: CommandIdentityCheckers（3 处内部参数类型） Co-Authored-By: glm-5.2 <zai-org@claude-code-best.win> * chore: 清理注释代码块与 legacy shim 注释代码（已死的、引用不存在符号的注释块）： - Onboarding.tsx: 注释化的 preflight if-block（引用不存在的 preflightStep） - ultraplan.tsx: 两处引用不存在符号的注释（ULTRAPLAN_INSTRUCTIONS、getUltraplanModel） - types/hooks.ts: 禁用的 type-fest IsEqual 类型断言块 - types/global.d.ts: 已被真实模块取代的 Ultraplan ambient declares - types/textInputTypes.ts: 注释化的 onMessage interface 成员 legacy shim： - cli/bg.ts: 删除 handleBgFlag 别名 export（同胞 handleBgStart 已被所有调用点使用） Co-Authored-By: glm-5.2 <zai-org@claude-code-best.win> * chore: 移除 ccshareResume stub 及 main.tsx 的 ccshare fast-path - 删除 src/utils/ccshareResume.ts（parseCcshareId 恒返回 null、loadCcshare 恒抛错的 stub） - 同步移除 src/main.tsx 中 USER_TYPE === 'ant' 守卫下的 if (ccshareId) {...} else {...} 双分支 - 提升 else 块（文件路径 resume 处理）为直接进入 if (options.resume) 块内 ccshare 是 Anthropic 内部特性（go/ccshare URL），stub 未实现导致 ccshareId 恒为 null，整个 ccshare 分支永不进入；保留的文件路径 resume 路径不变。 Co-Authored-By: glm-5.2 <zai-org@claude-code-best.win> * chore: 移除 environment-runner stub 及其 cli.tsx fast-path 与 self-hosted-runner 相同模式的 sibling（工作流 1 verifier 建议同步处理）： - 删除 src/environment-runner/main.ts（自动生成的 Promise.resolve() stub） - 同步移除 src/entrypoints/cli.tsx 中 feature('BYOC_ENVIRONMENT_RUNNER') 守卫的 fast-path 分支 - 清理两个空目录（src/self-hosted-runner/、src/environment-runner/） BYOC_ENVIRONMENT_RUNNER flag 不在 build.ts DEFAULT_BUILD_FEATURES 也不在 dev 默认列表，所有默认配置下整段为构建期死代码。 Co-Authored-By: glm-5.2 <zai-org@claude-code-best.win> * chore: 删除孤立诊断脚本 probe-local-wiring.ts #!/usr/bin/env bun shebang 的手动诊断脚本，全仓零引用，不在 package.json/build.ts/vite.config.ts/CI workflows 中。 Co-Authored-By: glm-5.2 <zai-org@claude-code-best.win> * chore: 移除 ultrareview preflight stub 及其测试 - 删除 src/services/api/ultrareviewPreflight.ts（自动生成的 stub） - 删除 src/commands/review/UltrareviewPreflightDialog.tsx（依赖前者的 UI stub） - 删除 src/services/api/__tests__/ultrareviewPreflight.test.ts（测试已删代码） - 同步移除 ultrareviewCommand.test.tsx 中对 UltrareviewPreflightDialog 的 mock Co-Authored-By: glm-5.2 <zai-org@claude-code-best.win> * chore: 移除 cachedMCConfig stub 及 prompts.ts 的 CACHED_MICROCOMPACT 死代码 - 删除 src/services/compact/cachedMCConfig.ts（自动生成的 stub） - 同步移除 src/constants/prompts.ts 中依赖该 stub 的代码： - getCachedMCConfigForFRC 变量（feature('CACHED_MICROCOMPACT') 守卫的 require） - getFunctionResultClearingSection 函数（约 18 行） - systemPrompt 数组中的 frc section 调用与注册 CACHED_MICROCOMPACT 不在 build.ts DEFAULT_BUILD_FEATURES 也不在 dev 默认列表，所有默认配置下整段为构建期死代码。 Co-Authored-By: glm-5.2 <zai-org@claude-code-best.win> * chore: 移除 goalAudit stub 及其测试引用 - 删除 src/services/goal/goalAudit.ts（导出 COMPLETION_AUDIT_RULES/BLOCKED_AUDIT_RULES/isGoalTerminal 等未引用的 stub） - 同步移除 tests/integration/goal-lifecycle.test.ts 中对 goalAudit 的 import 和一个测试用例（budget_limited is terminal） Co-Authored-By: glm-5.2 <zai-org@claude-code-best.win> * chore: 删除 agentSdkTypes 第二批 not-implemented stub 移除运行时函数体仅为 throw new Error 或 placeholder 的 stub： - createSdkMcpToolDefinition、createSdkMcpServer - query 函数重载与实现 - unstable_v2_* 系列函数 - session 操作 stub（getSessionMessages/listSessions/getSessionInfo/renameSession/tagSession/forkSession） - AbortError 类保留所有 export type 重导出和类型别名（仍是公共类型面）。 Co-Authored-By: glm-5.2 <zai-org@claude-code-best.win> * chore: 移除 Tool.ts 中 backwards-compat 重导出 shim 删除 "// Re-export progress types for backwards compatibility" 注释块及其重导出语句。所有消费方已直接从 src/types/tools.js 导入，无需重导出转发。 Co-Authored-By: glm-5.2 <zai-org@claude-code-best.win> * chore: 移除 bootstrap/state.ts 中 4 个未引用的 export - clearRegisteredHooks（STATE.registeredHooks 仍由其他函数管理） - getInvokedSkills（getInvokedSkillsForAgent 是活跃入口） - getSessionSource（setSessionSource 仍活跃，sessionSource state 字段保留） - markScrollActivity（scrollDraining/getIsScrollDraining/waitForScrollDrain 仍活跃）仅删除孤儿访问器，不动模块级 state 副作用。 Co-Authored-By: glm-5.2 <zai-org@claude-code-best.win> * chore: 移除 src/ 下多处未引用的导出涉及 18 个文件，每处均为独立的 unreferenced export 删除或 export 关键字冗余移除： - bridge/bridgeStatusUtil.ts、components/TrustDialog/utils.ts、context/stats.tsx - keybindings/loadUserBindings.ts、memdir/paths.ts、remote/sdkMessageAdapter.ts - services/acp/utils.ts（删除 nodeToWebReadable，全仓零引用） - services/api/metricsOptOut.ts、services/lsp/LSPDiagnosticRegistry.ts、services/lsp/manager.ts - services/mcp/utils.ts、services/skillLearning/projectContext.ts - services/teamMemorySync/secretScanner.ts、services/teamMemorySync/watcher.ts - skills/loadSkillsDir.ts、utils/attachments.ts、utils/filePersistence/filePersistence.ts - utils/messageQueueManager.ts Co-Authored-By: glm-5.2 <zai-org@claude-code-best.win> * chore: 移除 packages/ 下多处未引用的导出涉及 11 个 workspace 包文件，每处均为独立的 unreferenced export 删除或 export 关键字冗余移除： - @ant/ink/core/termio/csi.ts（eraseLine） - acp-link/manager/types.ts、acp-link/ws-message.ts - builtin-tools/AgentTool/agentMemory.ts、BashTool/bashSecurity.ts、BashTool/sedEditParser.ts - builtin-tools/ConfigTool/supportedSettings.ts、FileEditTool/utils.ts - remote-control-server/store.ts、transport/event-bus.ts、types/messages.ts Co-Authored-By: glm-5.2 <zai-org@claude-code-best.win> * Revert "fix: 终端内容溢出 viewport 时的重影 bug" This reverts commit 3d18e1da58. * revert: 移除主屏幕周期性 self-healing 重绘回退 f69c7051 中引入的 ink.tsx self-healing 机制（lastMainScreenHealTime 字段 + 每 5 秒触发全量重绘 + needsEraseBeforePaint 主屏幕分支）。该机制在 workflow 面板持续刷新场景下表现为可见的"重复刷新"，且修复效果不稳定。 alt-screen 的 needsEraseBeforePaint 路径和 prevFrameContaminated 字段保留，它们仍服务于 handleResize / layout shift / selection 高亮。 Co-Authored-By: glm-5.2 <zai-org@claude-code-best.win> * fix: /workflows 面板默认只显示运行中 run，根治 tab 行乱码之前几次渲染层修复都失败，因为没动 tab 列表的数据源：打开 /workflows 会自动 hydrate 最多 20 个历史 done/killed run，全部塞进一行 TabsBar，超出终端宽度后 Ink 把字符画到屏外造成重影乱码。 - selectors.ts 加 filterActiveRuns（只留 status === 'running'）和 capTabsForDisplay（超额 fold 成 +N）两个 pure function - WorkflowsPanel 接线 activeRuns：focus clamp、focused、nextTab/prevTab、 TabsBar 全部基于过滤后的 activeRuns - TabsBar 复用 truncateLabel 限制每个 tab 名 18 字符 + 最多 6 个 tab，多余显示 +N，从结构上钉死单行总宽度 Co-Authored-By: glm-5.2 <zai-org@claude-code-best.win> * fix: /workflows 面板 phase 状态在脚本省略 phase() 时显示错乱 ultracode canonical pipeline 脚本常在 agent() 直接传 opts.phase 而不调 phase() hook，导致 phase_started 从未发出；同时 phase_done 只在下次 phase() 触发，上一个 phase 在 run.phases 里一直停在 running。mergePhases 之前把 actual 当权威，于是出现 "Map 8/8 全 done 还显示 running、Find 1/4 running 反而显示 pending"。改为派生层修复：mergePhases 新增 derivePhaseStatus——actual.status==='done' 权威；否则有 agents 就按 agents 状态推（全 done→done，否则 running）；否则看 actual 是否 running。再补一层遍历，让只在 agents 上出现的 phase 也进 sidebar。不改 store 状态语义，已有 state.json 无需迁移。 Co-Authored-By: glm-5.2 <zai-org@claude-code-best.win> * docs: 更新 readme * fix: ACP 模式未读取 settings.local.json entry.ts 在 ACP 握手期调用的 applySafeConfigEnvironmentVariables 触发了 loadSettingsFromDisk，此时 getOriginalCwd() 还是进程启动 cwd（非项目目录），导致 localSettings/projectSettings 按错误路径解析为空并被 session cache 锁住，后续 createSession 里 setOriginalCwd 也无法纠正。在 setOriginalCwd 与 chdir 之后清缓存并重新应用，让 settings.local.json 和项目级 env 对 readSettingsPermissionMode 及下游可见。 Co-Authored-By: glm-5.2 <zai-org@claude-code-best.win> --------- Co-authored-by: glm-5.2 <zai-org@claude-code-best.win>
2026-06-25 17:45:50 +00:00 · 2026-06-22 09:59:36 +08:00
parent 8db85f1aaf
commit cd222b8e65
73 changed files with 304 additions and 2482 deletions
--- a/src/workflow/tests/selectors.test.ts
+++ b/src/workflow/tests/selectors.test.ts
@@ -2,6 +2,8 @@ import { expect, test } from 'bun:test'
 import type { AgentProgress, RunProgress } from '../progress/store.js'
 import {
  ALL_PHASE,
+  capTabsForDisplay,
+  filterActiveRuns,
  mergePhases,
  filterAgentsByPhase,
  tabLabel,
@@ -61,6 +63,57 @@ test('mergePhases: actual but undeclared phase appended to the end', () => {
  expect(mergePhases(r).map(p => p.title)).toEqual(['Find', 'Adhoc'])
 })

+// Regression: scripts that pass opts.phase directly to agent() without a phase() hook call
+// (the ultracode canonical pipeline pattern). phase_started is never emitted for those phases,
+// so run.phases lacks them. The sidebar used to show them as pending forever while agents were
+// clearly running under them — and worse, the previous phase stayed "running" because phase_done
+// only fires on the next phase() call. Derive status from agents when no actual record exists.
+test('mergePhases: derives status from agents when phase_started was never emitted', () => {
+  // Mirrors the real .claude/workflow-runs/wnxct9u3q/script.js shape:
+  // phase('Map') called, 8 Map agents done; pipeline stage with phase:'Find' running (1/4);
+  // Verify / Synthesize declared but not started; phase('Synthesize') not yet reached so
+  // phase_done Map has not fired either — actual Map is still 'running'.
+  const r = run({
+    declaredPhases: ['Map', 'Find', 'Verify', 'Synthesize'],
+    phases: [{ title: 'Map', status: 'running' }],
+    agents: [
+      ...Array.from({ length: 8 }, (_, i) => ({
+        id: i,
+        phase: 'Map',
+        status: 'done' as const,
+        resultKind: 'ok',
+      })),
+      { id: 100, phase: 'Find', status: 'done', resultKind: 'ok' },
+      { id: 101, phase: 'Find', status: 'running' },
+      { id: 102, phase: 'Find', status: 'running' },
+      { id: 103, phase: 'Find', status: 'running' },
+    ],
+  })
+  expect(mergePhases(r)).toEqual([
+    { title: 'Map', status: 'done', done: 8, total: 8 },
+    { title: 'Find', status: 'running', done: 1, total: 4 },
+    { title: 'Verify', status: 'pending', done: 0, total: 0 },
+    { title: 'Synthesize', status: 'pending', done: 0, total: 0 },
+  ])
+})
+
+// A phase that appears only on agents (not in declaredPhases, not in run.phases) is still
+// surfaced so the user sees it in the sidebar.
+test('mergePhases: phase only present on agents is appended and derived from agent states', () => {
+  const r = run({
+    declaredPhases: ['Scan'],
+    phases: [],
+    agents: [
+      { id: 1, phase: 'AdhocFromAgent', status: 'running' },
+      { id: 2, phase: 'AdhocFromAgent', status: 'done', resultKind: 'ok' },
+    ],
+  })
+  expect(mergePhases(r)).toEqual([
+    { title: 'Scan', status: 'pending', done: 0, total: 0 },
+    { title: 'AdhocFromAgent', status: 'running', done: 1, total: 2 },
+  ])
+})
+
 test('filterAgentsByPhase: All / undefined → all; specified → only that phase', () => {
  const agents: AgentProgress[] = [
    { id: 1, phase: 'A', status: 'running' },
@@ -80,3 +133,76 @@ test('filterAgentsByPhase: All / undefined → all; specified → only that phas
 test('tabLabel: workflow name + last 4 chars short code of runId', () => {
  expect(tabLabel('review-changes', 'wf_abc123def')).toBe('review-changes#3def')
 })
+
+// filterActiveRuns: only running runs reach the panel's tab row. Done/killed/completed are hidden
+// so opening /workflows no longer floods the tab row with months of historical runs (caused
+// tab overflow → garbled render when total width exceeded the terminal).
+test('filterActiveRuns: only status === "running" survives; completed/failed/killed dropped', () => {
+  const r1 = run({ runId: 'r1', status: 'running' })
+  const r2 = run({ runId: 'r2', status: 'running' })
+  const r3 = run({ runId: 'r3', status: 'completed' })
+  const r4 = run({ runId: 'r4', status: 'failed' })
+  const r5 = run({ runId: 'r5', status: 'killed' })
+  expect(filterActiveRuns([r1, r2, r3, r4, r5])).toEqual([r1, r2])
+})
+
+test('filterActiveRuns: empty input -> empty output', () => {
+  expect(filterActiveRuns([])).toEqual([])
+})
+
+test('filterActiveRuns: all terminal -> empty (panel falls back to "(no active runs)")', () => {
+  expect(
+    filterActiveRuns([run({ status: 'completed' }), run({ status: 'killed' })]),
+  ).toEqual([])
+})
+
+test('filterActiveRuns: preserves input order (no re-sort)', () => {
+  const a = run({ runId: 'a', status: 'running', startedAt: 5 })
+  const b = run({ runId: 'b', status: 'running', startedAt: 1 })
+  expect(filterActiveRuns([a, b]).map(r => r.runId)).toEqual(['a', 'b'])
+})
+
+// capTabsForDisplay: even if active runs somehow accumulate (long-lived sessions, runaway launcher),
+// the tab row must never overflow the terminal — cap at maxTabs, fold the remainder into a +N marker.
+test('capTabsForDisplay: under cap -> as-is', () => {
+  const runs = [
+    run({ runId: 'r1', status: 'running' }),
+    run({ runId: 'r2', status: 'running' }),
+  ]
+  expect(capTabsForDisplay(runs, 8)).toEqual({ runs, overflow: 0 })
+})
+
+test('capTabsForDisplay: over cap -> first maxTabs runs + overflow count', () => {
+  const runs = Array.from({ length: 10 }, (_, i) =>
+    run({ runId: `r${i}`, status: 'running' }),
+  )
+  const capped = capTabsForDisplay(runs, 8)
+  expect(capped.runs).toHaveLength(8)
+  expect(capped.runs.map(r => r.runId)).toEqual([
+    'r0',
+    'r1',
+    'r2',
+    'r3',
+    'r4',
+    'r5',
+    'r6',
+    'r7',
+  ])
+  expect(capped.overflow).toBe(2)
+})
+
+test('capTabsForDisplay: exactly at cap -> no overflow', () => {
+  const runs = Array.from({ length: 8 }, (_, i) =>
+    run({ runId: `r${i}`, status: 'running' }),
+  )
+  const capped = capTabsForDisplay(runs, 8)
+  expect(capped.runs).toHaveLength(8)
+  expect(capped.overflow).toBe(0)
+})
+
+test('capTabsForDisplay: maxTabs=0 -> all folded into overflow (degenerate but defined)', () => {
+  const runs = [run({ runId: 'r1', status: 'running' })]
+  const capped = capTabsForDisplay(runs, 0)
+  expect(capped.runs).toEqual([])
+  expect(capped.overflow).toBe(1)
+})
--- a/src/workflow/panel/TabsBar.tsx
+++ b/src/workflow/panel/TabsBar.tsx
@@ -3,21 +3,40 @@ import { Box, Text } from '@anthropic/ink';
 import type { Theme } from '@anthropic/ink';
 import type { RunProgress } from '../progress/store.js';
 import { RUN_STATUS_COLOR, STATUS_DOT } from './status.js';
-import { tabLabel } from './selectors.js';
+import { capTabsForDisplay, tabLabel } from './selectors.js';
+import { truncateLabel } from './AgentList.js';
+
+/**
+ * Per-tab name width budget. Long workflow names truncate (keeping the `#xxxx` short-code suffix so
+ * same-name runs stay distinguishable). Sized for a ~120-col terminal: ~6 tabs fit per row.
+ */
+const TAB_LABEL_MAX = 18;
+
+/**
+ * Hard ceiling on simultaneously rendered tabs. Defensive fallback: even if active runs accumulate
+ * (long-lived session, runaway launcher), the row must never overflow the terminal width and
+ * re-introduce the garbled overlapping render seen previously. Surplus runs are folded into `+N`.
+ */
+const MAX_TABS = 6;

 /**
 * Top run tab row: one tab per run (status dot + name + #short code).
 * The current tab is highlighted with an orange ═ underline.
+ *
+ * Defenses against overflow:
+ * - Per-tab name truncated via truncateLabel (keeps `#xxxx` suffix for disambiguation).
+ * - Row capped at MAX_TABS; remainder rendered as a `+N` marker so total width is bounded.
 */
 export function TabsBar({ runs, activeRunId }: { runs: RunProgress[]; activeRunId: string | null }): React.ReactNode {
  if (runs.length === 0) {
    return <Text color="subtle">(no runs)</Text>;
  }
+  const { runs: visible, overflow } = capTabsForDisplay(runs, MAX_TABS);
  return (
    <Box>
-      {runs.map(r => {
+      {visible.map(r => {
        const active = r.runId === activeRunId;
-        const label = tabLabel(r.workflowName, r.runId);
+        const label = truncateLabel(tabLabel(r.workflowName, r.runId), TAB_LABEL_MAX);
        const underline = '═'.repeat(label.length + 2);
        return (
          <Box key={r.runId} flexDirection="column" marginRight={2}>
@@ -32,6 +51,12 @@ export function TabsBar({ runs, activeRunId }: { runs: RunProgress[]; activeRunI
          </Box>
        );
      })}
+      {overflow > 0 ? (
+        <Box flexDirection="column" marginRight={2}>
+          <Text color="subtle">+{overflow}</Text>
+          <Text> </Text>
+        </Box>
+      ) : null}
    </Box>
  );
 }
--- a/src/workflow/panel/WorkflowsPanel.tsx
+++ b/src/workflow/panel/WorkflowsPanel.tsx
@@ -9,7 +9,7 @@ import { PhaseSidebar } from './PhaseSidebar.js';
 import { TabsBar } from './TabsBar.js';
 import { RUN_STATUS_COLOR, RUN_STATUS_TEXT } from './status.js';
 import { type FocusColumn, type WorkflowKeyboardHandlers, useWorkflowKeyboard } from './useWorkflowKeyboard.js';
-import { ALL_PHASE, filterAgentsByPhase, formatDuration, mergePhases } from './selectors.js';
+import { ALL_PHASE, filterActiveRuns, filterAgentsByPhase, formatDuration, mergePhases } from './selectors.js';

 /**
 * Clamp the selected index to a valid range (empty list -> 0; out of range -> last position; negative/NaN -> 0).
@@ -61,6 +61,10 @@ export function WorkflowsPanel({
    () => svc.listRuns(),
    () => [],
  );
+  // Only in-flight runs reach the tab row. Terminal (completed/failed/killed) runs are hidden so opening
+  // the panel no longer floods the row with persisted history (which overflowed the terminal and rendered
+  // garbled overlapping text). They stay on disk and remain resumable via getRunAsync.
+  const activeRuns = filterActiveRuns(runs);

  const [activeRunId, setActiveRunId] = useState<string | null>(null);
  const [focusColumn, setFocusColumn] = useState<FocusColumn>('phases');
@@ -76,18 +80,19 @@ export function WorkflowsPanel({
    void svc.loadPersistedRuns();
  }, [svc]);

-  // On runs change: activeRunId invalidated (killed / first time) -> clamp to the first one
+  // On activeRuns change: activeRunId invalidated (killed / first time) -> clamp to the first one.
+  // Tracks activeRuns (not raw runs) so focus never lands on a hidden terminal run.
  useEffect(() => {
-    if (runs.length === 0) {
+    if (activeRuns.length === 0) {
      if (activeRunId !== null) setActiveRunId(null);
      return;
    }
-    if (!runs.some(r => r.runId === activeRunId)) {
-      setActiveRunId(runs[0]!.runId);
+    if (!activeRuns.some(r => r.runId === activeRunId)) {
+      setActiveRunId(activeRuns[0]!.runId);
    }
-  }, [runs, activeRunId]);
+  }, [activeRuns, activeRunId]);

-  const focused: RunProgress | undefined = runs.find(r => r.runId === activeRunId);
+  const focused: RunProgress | undefined = activeRuns.find(r => r.runId === activeRunId);
  const phases = focused ? mergePhases(focused) : [];
  // The sidebar includes the All row: prepend one item to the phases array -> total rows = phases.length + 1
  const phaseRowCount = phases.length + 1;
@@ -122,15 +127,15 @@ export function WorkflowsPanel({
  };

  const nextTab = (): void => {
-    if (runs.length === 0) return;
-    const idx = runs.findIndex(r => r.runId === activeRunId);
-    const next = runs[(idx + 1) % runs.length]!;
+    if (activeRuns.length === 0) return;
+    const idx = activeRuns.findIndex(r => r.runId === activeRunId);
+    const next = activeRuns[(idx + 1) % activeRuns.length]!;
    switchTab(next.runId);
  };
  const prevTab = (): void => {
-    if (runs.length === 0) return;
-    const idx = runs.findIndex(r => r.runId === activeRunId);
-    const next = runs[(idx - 1 + runs.length) % runs.length]!;
+    if (activeRuns.length === 0) return;
+    const idx = activeRuns.findIndex(r => r.runId === activeRunId);
+    const next = activeRuns[(idx - 1 + activeRuns.length) % activeRuns.length]!;
    switchTab(next.runId);
  };

@@ -225,9 +230,9 @@ export function WorkflowsPanel({
      </Box>
      {focused?.description ? <Text color="subtle">{focused.description}</Text> : null}

-      {runs.length > 1 ? (
+      {activeRuns.length > 1 ? (
        <Box marginTop={1}>
-          <TabsBar runs={runs} activeRunId={activeRunId} />
+          <TabsBar runs={activeRuns} activeRunId={activeRunId} />
        </Box>
      ) : null}

--- a/src/workflow/panel/selectors.ts
+++ b/src/workflow/panel/selectors.ts
@@ -13,9 +13,40 @@ export type MergedPhase = {
 }

 /**
- * Merge declaredPhases (declared by meta) and run.phases (actually running/done):
- * - Declared order takes priority; phases present in actual but not declared are appended at the end.
- * - No actual record -> pending; otherwise take the actual status.
+ * Derive a phase's sidebar status from the actual record + the agents grouped under it.
+ *
+ * The actual record comes from `phase_started`/`phase_done` events. Scripts that follow the
+ * ultracode canonical pipeline pattern pass `opts.phase` directly to `agent()` inside
+ * `pipeline()`/`parallel()` stages and never call `phase()` for those phases — so no
+ * `phase_started` ever fires and `run.phases` lacks them. Worse, because `phase_done` only
+ * emits when the *next* `phase()` runs, the previous phase stays "running" in `run.phases`
+ * even after all its agents finish.
+ *
+ * Rules (checked in order):
+ * 1. `phase_done` already fired → done is authoritative, respect it.
+ * 2. Agents exist under this phase → derive from their states
+ *    (all done → done; otherwise → running). This is what the user actually sees.
+ * 3. No agents yet → fall back to the actual record
+ *    (`running` if `phase()` was called and is still active, else pending).
+ */
+function derivePhaseStatus(
+  actual: { status: 'running' | 'done' } | undefined,
+  inPhase: AgentProgress[],
+): PhaseStatus {
+  if (actual?.status === 'done') return 'done'
+  if (inPhase.length > 0) {
+    return inPhase.every(a => a.status === 'done') ? 'done' : 'running'
+  }
+  return actual?.status === 'running' ? 'running' : 'pending'
+}
+
+/**
+ * Merge declaredPhases (declared by meta), run.phases (actually running/done),
+ * and phases that appear only on agents:
+ * - Declared order takes priority; then actual-but-undeclared; then agent-only phases.
+ *   Agent-only phases surface in the sidebar even when the script never called `phase()`
+ *   for them — otherwise the user sees agents running under a phase that isn't listed.
+ * - Status is derived via {@link derivePhaseStatus}.
 * - done/total = done under that phase / total agents under that phase.
 */
 export function mergePhases(
@@ -28,17 +59,22 @@ export function mergePhases(
    if (seen.has(title)) return
    seen.add(title)
    const actual = actualByTitle.get(title)
-    const status: PhaseStatus = !actual ? 'pending' : actual.status
    const inPhase = run.agents.filter(a => a.phase === title)
    out.push({
      title,
-      status,
+      status: derivePhaseStatus(actual, inPhase),
      done: inPhase.filter(a => a.status === 'done').length,
      total: inPhase.length,
    })
  }
  for (const t of run.declaredPhases) push(t)
  for (const p of run.phases) push(p.title)
+  // Scripts that pass opts.phase directly to agent() (the ultracode pipeline pattern)
+  // may have agents grouped under phases that never got a phase() call — surface them
+  // so the sidebar reflects every phase the user can actually observe agents running in.
+  for (const a of run.agents) {
+    if (a.phase) push(a.phase)
+  }
  return out
 }

@@ -54,6 +90,37 @@ export function filterAgentsByPhase(
  return agents.filter(a => a.phase === selectedPhase)
 }

+/**
+ * Keep only runs still in flight. The /workflows panel defaults to this view: opening the panel
+ * no longer floods the tab row with months of persisted historical runs (which overflowed the
+ * terminal width and produced garbled overlapping text). Terminal runs (completed/failed/killed)
+ * stay on disk and remain resumable via getRunAsync; only the tab row filters them out.
+ *
+ * Pure + order-preserving: callers rely on the same relative order as the input (store.list()
+ * already returns newest-first by updatedAt).
+ */
+export function filterActiveRuns(runs: RunProgress[]): RunProgress[] {
+  return runs.filter(r => r.status === 'running')
+}
+
+/**
+ * Cap how many runs reach the tab row. Defensive fallback: even if active runs accumulate
+ * (long-lived session, runaway launcher), the row must never overflow the terminal width and
+ * re-introduce the garbled render. Anything past maxTabs is folded into an `overflow` count
+ * that the panel renders as `+N`.
+ *
+ * `runs` is sliced as-is (no re-sort); the caller is expected to have already applied
+ * filterActiveRuns and any ordering upstream.
+ */
+export function capTabsForDisplay(
+  runs: RunProgress[],
+  maxTabs: number,
+): { runs: RunProgress[]; overflow: number } {
+  const cap = Math.max(0, Math.trunc(maxTabs))
+  const visible = runs.slice(0, cap)
+  return { runs: visible, overflow: Math.max(0, runs.length - visible.length) }
+}
+
 /** tab label: workflow name + `#` + last 4 chars of runId (disambiguates same-name runs). */
 export function tabLabel(workflowName: string, runId: string): string {
  return `${workflowName}#${runId.slice(-4)}`