The prose gate reads an uppercase Markdown suffix as Markdown in one place and not in the others #2804

check_file in .github/actions/prose-gate/prose_lint.py compares path.suffix == ".md" case-sensitively at each of its Markdown-only branches, and comment_wrap_findings does the same for its unpunctuated-HTML-comment exemption. syntax_for lowercases the suffix, so a file named with an uppercase Markdown suffix has its comments extracted as Markdown but then misses every Markdown-only branch.

Constructed case: a file NOTES.MD holding <!-- group header --> draws comment-case, where the same content in NOTES.md is exempt as a structural marker. The sentence-split, sentence-length, dead-path and quote-stripping branches are skipped for it entirely.

The .py comparison in comment_wrap_findings was made case-insensitive on the branch fixing #1659, which is where this was found by a local review pass. The .md comparisons were left alone there since they sit outside that issue. The fix is presumably one lowercased suffix read once per call and used at every comparison, with a case pinning an uppercase .MD against each branch.