[Feature]: Improve skip onboarding UX to communicate incomplete setup

[HeyItsChloe]

Is there an existing feature request for this?

• I have searched existing issues and feature requests, and this is not a duplicate.

Problem or Use Case

When a user clicks "Skip for now" during onboarding, the system only sets a localStorage flag (openhands-onboarded) but saves NO settings to the agent-server. The user lands on the conversation page with no indication that the conversations will fail.

Current Behavior:

• "Skip for now" calls markCompleted() which sets localStorage["openhands-onboarded"] = "1"
• No settings are persisted to agent-server
• User sees normal home page UI
• Any conversation attempt fails silently

Proposed Solution

Implement one or more of these options:

Option A: Show persistent banner when LLM not configured

• Display warning banner on home page: "LLM not configured. Set up your model to start conversations."

Option B: Redirect to settings after skip

• After skip, redirect to /settings/llm with a message

Option C: Disable new conversation UI

• Gray out "New Conversation" button with tooltip explaining setup needed

Alternatives Considered

• Removing the "Skip for now" button entirely - rejected because users may want to explore the UI first
• Auto-configuring a default LLM - rejected because it requires API key which user must provide

Priority / Severity

High - Significant impact on productivity

Estimated Scope

Medium - New feature with moderate complexity

Feature Area

User Interface / UX

Technical Implementation Ideas

Files to Modify:

• src/components/features/onboarding/onboarding-modal.tsx
• src/components/features/onboarding/use-onboarding-completion.ts
• src/routes/home.tsx

Could add a new hook useLlmConfigured() that checks llm_api_key_is_set and use it to conditionally render warning UI.

Acceptance Criteria

• User understands system is not ready after skipping onboarding
• Clear path to complete LLM setup is provided
• No silent failure when attempting to use the system

---

This issue was created by an AI agent (OpenHands) on behalf of the user.