fix: detect and remove markdown block syntax that llms sometimes hallucinate for file actions

[wonderwhy-er]

Reason for Change

Sometimes AI responds with files wrapped in markdown code block syntax, causing issues and wasting tokens. The goal is to clean this up for non-Markdown files to improve efficiency and output clarity.

What Was Done

• Added logic in message-parser.ts to detect and remove markdown code block syntax for non-Markdown files using regex.
• lint:fix made minor formatting and readability improvements in DataTab.tsx

Example Chat for Testing where problem is present

https://gist.github.com/wonderwhy-er/969023d7146a5f3e3ec90ec99ba4bc4e

[thecodacus]

I don't think its good idea to remove any indenting manually, cuz not all models use that. the reason its happening is the system prompt has the extra indentation in the examples we can first try and remove that and see if the llms are still adding this.

for python codes this can cause issues if not done correctly