fix: prevent # symbols in code blocks from being treated as headers in explode command

[halso]

Summary

• Fixes issue Markdown Tree Parser Truncates Sections At # Comments Within Embedded Bash Code Blocks #7 where the explode command would incorrectly truncate sections when encountering # symbols inside code blocks
• Now properly tracks code block boundaries to skip header detection when inside code blocks
• Includes comprehensive test coverage for this scenario
• Also includes automatic formatting/linting changes for code consistency

Problem

The explode command was treating # symbols inside code blocks (like bash comments) as markdown headers, causing sections to be truncated prematurely.

Solution

Added code block state tracking:

1. Detect code block boundaries (```)
2. Skip header detection when inside code blocks
3. Only process lines as headers when NOT in a code block

Commits in this PR

1. chore: apply automatic formatting/linting changes and add CLAUDE.md to gitignore

   • Applied Prettier formatting to JSON and markdown files (2-space indentation)
   • Added newline at end of index.js
   • Manually added CLAUDE.md to .gitignore for AI assistant workspace analysis

2. fix: prevent # symbols in code blocks from being treated as headers in explode command

   • Implemented the actual bug fix for issue Markdown Tree Parser Truncates Sections At # Comments Within Embedded Bash Code Blocks #7
   • Added test coverage for the scenario

Test plan

• All existing tests pass (npm test)
• Linting passes (npm run lint)
• Code is formatted (npm run format)
• CLI tests pass (npm run test:cli)
• Added specific test case for issue Markdown Tree Parser Truncates Sections At # Comments Within Embedded Bash Code Blocks #7 that validates the fix

Closes #7

🤖 Generated with Claude Code

[ksylvan]

A lot of this seems like formatting changes - please reduce this diff to just the functional changes and keep my formatting choices.

Also, please squash your commits into 1.

[halso]

@ksylvan - you're right. They shouldn't be mixed in. The formatting changes were done by the tools automatically when I ran your instructions at:

https://github.com/ksylvan/markdown-tree-parser/blob/main/CONTRIBUTING.md#before-submitting

so I wanted to keep them separate from the actual bug commit.

I'm guessing the correct thing to do then is:

• You or I create an issue for the fact the auto-formatting has changed and commit the changes (let me know if you want me to do it).
• I create an issue for the "chore" commit of adding the .gitignore for CLAUDE.md and submit a pull request.
• I then just retract my mixed up pull request (or you reject it) and I submit a clean fix on this Issue with just the changes to test and the src code.

Sorry for the confusion. It's my first time doing a Pull Request and also wasn't sure what the best thing was to do (and didn't want to create lots of different GitHub issues).

[xymb-endcrystalme]

A lot of this seems like formatting changes - please reduce this diff to just the functional changes and keep my formatting choices.

Maybe just apply the first commit of this pull request: c32cfad
Or should I open a pull request with only the first commit?

The markdown-tree-parser is definitely buggy without this fix.

[ksylvan]

I'll take care of this today. Thank you for finding the bug.

[ksylvan]

Fixed in #9