fix: use llmMessage JSON schema in prompt

[codeincontext]

On a local branch I accidentally disabled structured outputs and I found that the LLM message was in the wrong structure. Specifically, the "prompt" key in the llmMessage type was this shape:

  message: "Are the learning outcome and learning cycles appropriate for...",
  type: "text"

When it should have been

  value: "Are the learning outcome and learning cycles appropriate for...",
  type: "text"

This is because we were providing the schema for old parts in the prompt:

  Patch
  Prompt
  State
  Comment
  Error

Because there wasn't a schema for LLMMessage, it tried the structure of the Prompt part

Description

• Replace old part types in the schema with the LLMMessage part schema

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name	Status	Preview	Comments	Updated (UTC)
oak-ai-lesson-assistant	✅ Ready (Inspect)	Visit Preview	💬 Add feedback	Mar 6, 2025 9:19am

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[github-actions]

Playwright test results

 17 passed
 2 skipped

Details

 Open report ↗︎
 19 tests across 16 suites
 2 minutes, 36 seconds
 82394ac

Skipped tests

No persona › tests/auth.test.ts › authenticate through Clerk UI
No persona › tests/chat-performance.test.ts › Component renders during lesson chat › There are no unnecessary rerenders across left and right side of chat