feat(cli): simplify auto mode startup text and remove emoji (#4584)

[ZijianZhang989]

What this PR does

Removes the ✨ emoji prefix from the Auto mode first-time startup message and simplifies the copy from 5 lines to 3 lines. The message now reads: "Auto mode enabled. An LLM classifier evaluates each tool call — safe actions auto-approve, risky ones are blocked. Exit: Shift+Tab or /approval-mode default."

Why it's needed

The emoji icon in front of the auto mode startup text looked visually redundant, and the original copy was unnecessarily verbose. This simplifies the text to be more concise and professional while preserving all essential information (what auto mode does, how to exit).

Resolves #4584.

Reviewer Test Plan

How to verify

cd packages/cli && npx vitest run src/ui/hooks/useAutoAcceptIndicator

Manually: start the CLI with --approval-mode auto on a fresh config (no ui.autoModeAcknowledged set) and confirm the message appears without an emoji prefix and with shortened text.

Evidence (Before & After)

Before:

✨ Auto mode enabled.
   An LLM classifier evaluates each tool call and auto-approves safe actions,
   blocks risky ones. Most read-only operations and in-cwd edits skip the
   classifier for speed. To exit: Shift+Tab or /approval-mode default.
   (This notice will not appear again.)

After:

Auto mode enabled.
   An LLM classifier evaluates each tool call — safe actions auto-approve,
   risky ones are blocked. Exit: Shift+Tab or /approval-mode default.

Tested on

OS	Status
🍏 macOS	⚠️
🪟 Windows	⚠️
🐧 Linux	⚠️

Risk & Scope

• Purely cosmetic text change — no logic modified
• Breaking changes: None

Linked Issues

Closes #4584

中文说明

这个 PR 做了什么

移除 Auto 模式首次启动提示文字前的 ✨ emoji，并将文案从 5 行精简为 3 行。新文案为："Auto mode enabled. An LLM classifier evaluates each tool call — safe actions auto-approve, risky ones are blocked. Exit: Shift+Tab or /approval-mode default."

为什么需要

原有的 emoji 图标在文字前面显得多余，原始文案也过于冗长。简化后更加简洁专业，同时保留了所有关键信息（auto 模式的作用、退出方式）。

解决 #4584。

Reviewer Test Plan

手动验证：用 --approval-mode auto 启动 CLI（清除 ui.autoModeAcknowledged 设置），确认消息无 emoji 前缀且文案已缩短。

风险与范围

• 纯文案变更，无逻辑修改
• 破坏性变更： 无

关联 Issues

Closes #4584

[wenshao]

No issues found. LGTM! ✅

— qwen3.7-max via Qwen Code /review

[qwen-code-ci-bot]

[Suggestion] Cross-surface string drift: packages/web-shell/client/i18n.tsx (lines 777 and 1845) still uses the old phrasing — "auto-approves safe actions, blocks risky ones" — while the CLI message now reads "safe actions auto-approve, risky ones are blocked". Consider updating the web-shell copy (both English and Chinese locales) in this PR to keep the two surfaces consistent.

[Nice to have] The stripped-rules notice at line 169 still uses the ℹ️ emoji ('ℹ️ Auto mode temporarily disabled these allow rules'), while the first-time message is now emoji-free. Since both messages can appear back-to-back on first AUTO entry, consider removing the emoji there too for visual consistency.

— qwen3.7-max via Qwen Code /review

[qwen-code-ci-bot]

Thanks for the PR, @ZijianZhang989!

Template looks good ✓

On direction: this is a straightforward cosmetic improvement — removes the ✨ emoji prefix from the auto mode first-time message and tightens the copy. Solves a real annoyance (verbose startup text) and aligns with issue #4584. Auto mode UX is well-represented in competitor changelogs, so polishing it is worthwhile.

On approach: the scope is exactly right — 1 file, 1 constant, pure text substitution. No scope creep, no drive-by refactors. The three removed details (emoji, "Most read-only operations skip the classifier" implementation note, "This notice will not appear again" meta-comment) are all reasonable cuts: the emoji was decorative, the classifier detail was internal implementation info users don't need, and the "won't appear again" note is self-evident after the first session.

Moving on to code review. 🔍

中文说明

感谢贡献，@ZijianZhang989！

模板完整 ✓

方向：这是一个直接的 UI 改善——移除 auto 模式首次启动提示前的 ✨ emoji 并精简文案。解决了 #4584 提出的问题。Auto 模式的 UX 在竞品 changelog 中也有多处提及，打磨它是有价值的。

方案：范围恰好——1 个文件、1 个常量、纯文本替换。没有夹带无关改动。删除的三处内容（emoji、"大多数只读操作跳过分类器"的实现细节、"此提示不再出现"的元注释）都是合理的裁剪。

进入代码审查 🔍

— Qwen Code · qwen3.7-max

[qwen-code-ci-bot]

Code Review

Independent proposal before reading the diff: I would change the AUTO_MODE_FIRST_TIME_MESSAGE constant in useAutoAcceptIndicator.ts — strip the emoji, condense the multi-line message into 2-3 lines, keep the essential info (what auto mode does, how to exit). Remove the "won't appear again" meta-note since it's self-evident.

The PR matches this exactly. The diff is a clean 5-line deletion / 3-line addition in a single string constant. No logic changes, no new imports, no test updates needed (no test file exists for this hook, and no existing test references the message text). The em-dash restructuring ("— safe actions auto-approve, risky ones are blocked") reads better than the original run-on sentence.

No correctness bugs, no security concerns, no convention violations. This is about as minimal as a PR can get.

Real-Scenario Testing

Before (installed build, main branch)

  Tips: Type / to open the command popup; Tab autocompletes slash commands and saved prompts.
  ● ✨ Auto mode enabled.
       An LLM classifier evaluates each tool call and auto-approves safe actions,
       blocks risky ones. Most read-only operations and in-cwd edits skip the
       classifier for speed. To exit: Shift+Tab or /approval-mode default.
       (This notice will not appear again.)

After (this PR, dev build)

  Tips: Type / to open the command popup; Tab autocompletes slash commands and saved prompts.
  ● Auto mode enabled.
       An LLM classifier evaluates each tool call — safe actions auto-approve,
       risky ones are blocked. Exit: Shift+Tab or /approval-mode default.

Before/after confirms: emoji removed ✓, text condensed from 5 lines to 3 ✓, essential info preserved (what auto mode does, exit shortcut) ✓. No regressions visible — the TUI renders correctly with the shorter message.

中文说明

代码审查

独立方案：修改 useAutoAcceptIndicator.ts 中的 AUTO_MODE_FIRST_TIME_MESSAGE 常量——去掉 emoji，将多行消息压缩为 2-3 行，保留关键信息。

PR 完全符合预期。diff 是一个干净的单文件字符串常量修改，无逻辑变更、无新导入。em-dash 重构（"— safe actions auto-approve, risky ones are blocked"）比原来的长句更清晰。

无正确性问题、无安全隐患、无规范违反。

真实场景测试

前后对比确认：emoji 已移除 ✓，文案从 5 行精简到 3 行 ✓，关键信息保留（auto 模式功能、退出快捷键）✓。TUI 渲染正常，无回归。

— Qwen Code · qwen3.7-max

[qwen-code-ci-bot]

This is a clean, minimal cosmetic change that does exactly what it promises. The before/after tmux output confirms the emoji is gone, the text is tighter, and nothing important was lost. The PR author made good editorial choices — cutting the implementation-detail sentence and the meta "won't appear again" note while keeping the actionable info (what auto mode does, how to exit).

My independent proposal was identical to what the PR does. No simpler path exists — it's already a single-constant edit. The diff carries zero unrelated changes.

LGTM, ships the improvement cleanly. ✅

中文说明

这是一个干净、最小化的 UI 改动，完全兑现了 PR 描述。前后对比确认 emoji 已移除、文案更精简，且没有丢失重要信息。作者在编辑取舍上做得好——删除了实现细节和元注释，保留了可操作的信息。

我的独立方案与 PR 完全一致，没有更简的路径。diff 零夹带改动。

可以合入 ✅

— Qwen Code · qwen3.7-max

[qwen-code-ci-bot]

LGTM, looks ready to ship. ✅