[LIHUA919]

Goal And Delivered Outcome

• Outcome basis / optional anchor: the required dashboard-browser and chat-bundle-browser lanes, red on main at f5e9be367. They fail at goal-suggestions-live.mjs:65 with "the lane itself shows the item as optional", and the vitals text shows 状态 0/0 工作 0 项未完成 暂无等待 等你决定.
• Goal/source and gap: the scenario adds an optional user_action Todo through the real CLI after the page has already loaded, then reads the Goal vitals without refreshing status. The view still shows the empty Goal. The old assertion /0 等你决定|可选/ passed only because the empty needs-you cell used to render a bare 0. fix(dashboard): readable Explore thirds and Goal empty states #7478 changed that cell to read 暂无等待, which exposed the stale read. A first-parent bisect over 5a65338fa..f5e9be367 names adfcfb5a2 (fix(dashboard): readable Explore thirds and Goal empty states #7478) as the first failing commit.
• Fix: refresh status after adding the item, as the user-gate step a few lines later already does. Then assert the optional count the lane really renders (另有 1 项可选).
• Observable before → after: the scenario fails at the main head before this change and passes after it. The "optional item is listed, but not as a decision" check now proves the item is actually loaded.
• Issue/task and intended base: related to CI: chat-bundle-browser and dashboard-browser shards time out on main #7389. Follows fix(dashboard): readable Explore thirds and Goal empty states #7478. Base main.

Author Declaration

• Written by: model_agent (Claude Opus, Anthropic), at the account owner's direction.

Implemented against

• Specification and revision: the scenario's own header contract ("An optional user action raised through the CLI is listed but never becomes a 'decision' suggestion") and vitals.needsYouOptional (另有 {count} 项可选).

Criterion (spec clause)	Disposition	Symbol / path	Test or command
The optional item is read after it is listed	implemented	examples/personal-workspace-browser/goal-suggestions-live.mjs	LOOPX_PERSONAL_WORKSPACE_SCENARIO=goal-suggestions-live
The lane shows it as optional, not a decision	strengthened	same	same

• Self-check before submission: negative control. Without the added refresh(), the new assertion fails with the same stale 0/0 vitals that main shows.

Scope And Continuation

• Completed scope: test only. No product code changed.
• Not covered: on the same main head, kernel-contracts and windows-powershell fail the CLI output budget. quota_should_run/multi_agent/json emits 25,836 characters against a 25,500 ceiling. That growth comes from Preserve complete context in shared quota diagnostics #7557, because the budget smoke passes at 5a7cc8195. Choosing a ceiling or compaction belongs to the quota owner, so this PR leaves it alone.

Validation

• Tested revision: f5ed559e0 on main f5e9be367 (rebased without conflicts to e5e16a414 on 1a7234c2a).
• Run state: finished (local); hosted CI pending on this PR.
• Input classes: synthetic

Check kind	Result	Public-safe evidence / limitation
real_entrypoint	passed	goal-suggestions-live: 3/3 in development mode and 2/2 packaged, at main f5e9be367.
regression_parity	passed	The main head fails before the change. Without the refresh, the new assertion fails.

Frontend / Visual Evidence

• UI impact: none (test only).

Type of Change

• Test update

LoopX Area

• Build, packaging, installer, or CI

Shared-authority RFC fixture impact

N/A

Boundary Checklist

• Neither the diff nor this PR body/comments/attachments disclose private state, credentials, raw traces or verifier output, internal links, or local machine paths.
• I did not duplicate maintainer-owned benchmark work.
• I kept the change scoped to the linked issue/task.
• No paired RFC changed.
• UI impact none.
• Every commit includes a DCO Signed-off-by trailer.

🤖 Generated with Claude Code

[loopx-agent]

Reviewer: model_agent; gpt-6.1-sol; OpenAI; runtime_reported; reasoning_effort=max

未发现阻塞问题，批准准确 head e5e16a4148fd29ef93e0a35635b3e179932e977c。

动机 / Motivation

运行 Dashboard 与 Chat browser smoke 的维护者需要确认 CLI 新增的可选事项确实被页面读取。以前页面仍是空 Goal 就检查可选状态；现在先刷新，再要求实际出现一项可选事项。旧脚本及移除刷新后的负控都失败，新 head 的开发态和打包态沿真实 Todo/status 路径通过。本 PR 只修现有测试，不改变产品页面、自动刷新策略、Todo 权限或其它主干问题。

改动思路 / Approach

沿用 user_gate 步骤已经使用的 refresh()，让普通页面读取真实新增项；再把允许空状态混过的正则收紧为实际可选计数。不增加 sleep、轮询器、场景或产品刷新规则。只改一处既有 smoke，边界足够完整且容易回退。

具体改动 / Changes

依据 examples/personal-workspace-browser/goal-suggestions-live.mjs 在 spec_revision 1a7234c2afcc3fd7897b63d5bab78d3e6a9f7959 的原头部合同：optional_not_decision 要求 CLI 可选事项确实列出、但不成为决定；gate_uses_owner_count 要求真正 gate 的建议与 typed lane 同数并打开原 gate。两项在本 head 都执行通过。

关键内容讲解

optional 段落 在 CLI add 成功后刷新；随后第 66 行断言“另有 1 项可选”，而不是仅满足空的零决定。后半段真实 gate 的 count equality 与导航断言保留。实际源是临时真实 Todo/status backend，只有 status 路径被代理，未把期望计数直接塞给组件。

对主干的风险 / Risks

PASS：精确 head 的开发态一次、重新构建后的打包态一次，完整走到 gate 导航；Dashboard/Chat build 和 standard premerge 的 4 个直接、3 个选中检查均通过，0 hold。保留 Vite 的大 chunk advisory warning。另在完全相同产品源码下执行不可变 base 的原场景，以及 head 仅移除首次刷新后的 mutation：两者都在可选计数处失败，呈现同一空 Goal，而新 head 能读到可选项。没有删除 gate 断言、放宽预算或隐藏失败，也未查询 CI。

对比既有纯 fixture goal-suggestions、同作者近期 #7554 折叠组、#7550 来源返回、#7569 CLI 参数测试，此处修复的是不同的真实读取边界；没有新增重复 smoke，保留有实际维护价值。

我的整体评价 / Verdict

这个测试修复达成原场景目标：持续回归的可信度提高，产品交互保持原状。测试没有改变渲染组件；检查了 goal-suggestions-live.mjs 与 goal-masthead.tsx 的既有可选/阻塞计数路径。未据测试通过宣称视觉质量、正式 App 采用、整个 browser shard 或持续时序稳定已验收。普通 fixture 不能替代真实 CLI 新增后的读回，因此现有 live case 应保留；无需额外架构或跟进任务。批准源码与测试结论，合并资格另由当前 native closeout/readiness 判断。

English verdict: APPROVE e5e16a4148fd29ef93e0a35635b3e179932e977c; the real optional-item read passes in development and packaged modes, while immutable-base and no-refresh controls fail as expected.