fix: filter unlisted pages from llms.txt when sitemap fallback is used

[prajakta128]

This PR fixes a bug where pages marked unlisted: true in frontmatter were never filtered out of generated llms.txt files when the sitemap fallback path was used.

The issue was caused because buildReplacementsAndUnlisted() correctly returned both replacements and unlistedPaths, but main() only extracted replacements. As a result, the computed unlistedPaths set was discarded before it could be used during pruning.

Problematic code:

const { replacements } = buildReplacementsAndUnlisted();

Fixed code:

const { replacements, unlistedPaths } =
  buildReplacementsAndUnlisted();

This bug is silent in production environments where sitemap.xml exists, but it becomes active in local development and partial CI builds that rely on the globalData.json fallback path.

Changes included:

• Threaded unlistedPaths through main()
• Updated pruneAndMarkdownify() signature to accept unlistedPaths
• Added explicit unlistedPaths.has(route) filtering logic
• Updated processLlmsTxt() signature and call sites to pass unlistedPaths

No new dependencies added. No behavior change when sitemap.xml is present.

[prajakta128]

Hi,

I have submitted PR #1401 for review. This fix addresses
a bug where unlisted pages were not being filtered from
llms.txt when the sitemap fallback path was used.

Could you please:

1. Approve the pending workflow to allow the CI checks to run
2. Review and approve the pull request

Summary of changes:

• Threaded unlistedPaths through axis()
• Updated processMarkdownly() to accept unlistedPaths
• Added explicit unlistedPaths.has(route) filtering logic
• Updated process.Init() signature to pass unlistedPaths

No new dependencies have been added. There is no behavior
change when sitemap.xml is present.

PR: #1401

Please let me know if you have any questions or require
any changes.

Thank you.