convertAllMdx: add a `prune` option to remove orphaned outputs on rename/delete

[poiley]

Summary

convertAllMdx writes a .md file into outDir for each source page, but it never removes outputs whose source was deleted or renamed. Over time outDir accumulates orphaned files that no longer correspond to any source page, and they keep getting served/published.

Repro

1. Author guides/old-thing.mdx, run convertAllMdx → outDir/guides/old-thing.md is written.
2. Delete or rename guides/old-thing.mdx, run convertAllMdx again.
3. outDir/guides/old-thing.md is still there. It's now a dead URL with stale content, and it leaks into anything that walks outDir (sitemaps, link checks, search indexing of the raw files, etc.).

Why it matters

Renames are common during docs work, and a stale .md is worse than a 404 — it's a live page with outdated content and no source of truth. Every consumer that wants a clean outDir has to implement the same garbage-collection step: compute the expected output set from the resolved pages, walk outDir, and delete any .md not in the set. That's easy to get wrong (e.g. accidentally deleting non-generated files that share the directory).

Current workaround (illustrative)

const expected = new Set(pages.map((p) => path.join(outDir, p.markdownPath)))
for (const file of await walk(outDir)) {
  if (file.endsWith('.md') && !expected.has(file)) await rm(file, { force: true })
}

Proposed change

Add an opt-in prune option to convertAllMdx (default false to preserve current behavior):

await convertAllMdx({ srcDir, outDir, prune: true /* , remarkPlugins */ })

When prune: true, after writing the current set of outputs, remove any .md under outDir that wasn't produced by this run. leadtype already knows the authoritative set of output paths, so it can prune safely (and scope strictly to files it owns) in a way a consumer can't do as reliably from outside. Happy to PR.

[coderabbitai]

🔗 Related PRs

#23 - Polish FileTree, Prompt, and TopicSwitcher in docs example [merged]
#32 - Shrink leadtype dependency closure [merged]
#35 - Add v1 headless integration surface and polish docs [merged]
#106 - Use Satteri for MDX parsing [merged]
#112 - Batch git frontmatter enrichment [closed]

---

📝 Issue Planner

_{Check the box below or use the @coderabbitai plan command to generate an implementation plan and prompts that you can use with your favorite coding assistant.}

• Create Plan

---

🧪 Issue enrichment is currently in open beta.

You can configure auto-planning by selecting labels in the issue_enrichment configuration.

To disable automatic issue enrichment, add the following to your .coderabbit.yaml:

issue_enrichment:
  auto_enrich:
    enabled: false

💬 Have feedback or questions? Drop into our discord!

[pullfrog]

  ^{｜ Build this ➔ ｜ Make a plan ➔}