[UX Enhancement] Add configurable visual separator between AI responses for improved readability

Preflight Checklist

• [x] I have searched existing requests and this feature hasn't been requested yet
• [x] This is a single feature request (not multiple features)

Problem Statement

xxxxxxx

Proposed Solution

...xxxxxxxxx

Alternative Solutions

...xxxxxxxxxx

Priority

Critical - Blocking my work

Feature Category

CLI commands and flags

Use Case Example

...xxxxxxxxx

Additional Context

GitHub Issue - Claude Code Visual Separator

INFORMAÇÕES PARA PREENCHIMENTO

URL para criar issue: https://github.com/anthropics/claude-code/issues/new

Título da issue:

[UX Enhancement] Add configurable visual separator between AI responses for improved readability

Labels sugeridas:

• enhancement
• ux
• accessibility

---

CORPO DA ISSUE (copiar e colar)

## 🎯 Problem Statement

During extended Claude Code sessions, it becomes challenging to visually distinguish where one AI response ends and another begins, especially when:
- Scrolling back through conversation history
- Experiencing visual fatigue during long coding sessions
- Working with multiple consecutive responses
- Reviewing past interactions to understand context

**Current behavior:** The CLI automatically collapses all blank lines and markdown separators (like `---`) at the beginning of responses, resulting in responses appearing visually concatenated.

## 📸 Current Experience

question 1

● Response 1 text here...

question 2

● Response 2 text here...

*(No clear visual boundary between responses)*

## ✨ Proposed Solution

Add a **user-configurable setting** for visual separation between responses. Possible implementations:

### Option 1: Preserved Blank Lines (Simplest)
```json
// .claude/settings.json
{
  "ui.responseSeparator": {
    "blankLines": 3  // Number of blank lines preserved at start of response
  }
}

Option 2: Custom Separator Pattern

{
  "ui.responseSeparator": {
    "enabled": true,
    "pattern": "═══════════════════════════════════════════════",
    "color": "dim"  // Optional: terminal color code
  }
}

Option 3: Auto-timestamp Header

{
  "ui.responseSeparator": {
    "showTimestamp": true,  // Displays "[22:43:24] ──────"
    "format": "HH:mm:ss"
  }
}

🎨 Expected Behavior (Example with Option 2)

> question 1
● Response 1 text here...

═══════════════════════════════════════════════
> question 2
● Response 2 text here...

💼 Use Cases

1. Accessibility: Users with visual processing difficulties need clear boundaries
2. Long sessions: Developers in extended debugging/refactoring sessions
3. Context switching: Quickly scanning conversation history to find specific exchanges
4. Documentation: Creating session logs that are readable when exported

🔧 Current Workaround

Users must manually execute clear/cls between interactions, which:

• ❌ Disrupts workflow
• ❌ Loses conversation history visibility
• ❌ Requires manual action after every response

📊 Impact

• Severity: Low/Medium (quality of life improvement)
• Frequency: Affects every session
• User base: Benefits all users, critical for those with accessibility needs

🙋 Reporter Context

Reported by Rafael Silva
User experience feedback from production usage

---

Note: I attempted to implement this via custom markdown formatting but the CLI's processing layer strips all spacing/separator attempts. A configuration-level solution would properly address this UX gap.

---

## PASSOS PARA POSTAR

1. Abra no navegador: https://github.com/anthropics/claude-code/issues/new
2. Faça login no GitHub se necessário
3. Cole o TÍTULO no campo "Title"
4. Cole o CORPO DA ISSUE no campo de texto grande
5. Adicione as labels sugeridas (se disponível)
6. Clique em "Submit new issue"

---

## OPCIONAL: APÓS POSTAR

- Compartilhe o link da issue no Discord da Anthropic: https://anthropic.com/discord
- Mencione no canal #claude-code ou #feedback
- Aumenta visibilidade e engajamento da comunidade