[Feature] Add LLM provider setup wizard and model selector UI

[d-oit]

Description

The AI Harness requires API keys to function, but there is no setup wizard or guidance in the UI. Users must manually set environment variables or localStorage values. Additionally, there is no model selector — models are hardcoded in the provider code.

Current State

• ❌ No setup wizard for API keys
• ❌ No model selector UI
• ❌ No rate limiting or token usage display
• ❌ Raw LLM output rendered without Markdown formatting
• ❌ onKeyPress is deprecated (should be onKeyDown)
• ❌ No error messages when API key is missing

Proposed Implementation

1. Setup wizard: On first visit to AI Harness, guide user to enter API key and select provider
2. Model selector: Dropdown with available models per provider (OpenRouter, Kilo, etc.)
3. Token usage: Display tokens used per message and per session
4. Rate limiting: Visual feedback when rate-limited (429 responses)
5. Markdown rendering: Render LLM responses with Markdown formatting
6. Fix deprecated onKeyPress → onKeyDown

Acceptance Criteria

• Setup wizard guides API key configuration
• Model selector dropdown available
• Token usage displayed per message/session
• LLM responses rendered as Markdown
• onKeyPress replaced with onKeyDown
• Error messages for missing/invalid API keys

[github-actions]

👋 This looks like a duplicate of #183.

The canonical issue has broader scope and/or more tasks — please subscribe there for updates.

⏳ This issue will be auto-closed in 3 days unless you react with 👎 to dispute.

Detected via title similarity analysis