[Birfy]

Owner: Dev
Reviewer: Reviewer
Written-by: Dev

Fixes #909
Refs #910

Tip SHA: e1f924a

#910 merged without a changelog.d/ fragment, so this publishes the security fix it shipped: one paragraph, named after issue #909, in .changelog.d/. No source, no docs, no behaviour — the fragment is the whole diff (+1/−0).

Evidence on e1f924a: node .github/scripts/changelog.mjs check → 46 fragments well formed. Nothing else ran; changelog.d/ is not owned by a subsystem doc, so arch-docs has nothing to ask for.

This is changelog.d/* only — not the mechanical release PR's file set (CHANGELOG.md + version), so it takes the ordinary path: green tests-ok, mergeable, merged by whoever gates it.

Not run: the test suites (no code changes), arch-docs (needs a git checkout; no docs/subsystems file changed).

[Birfy]

Verdict: PASS at e1f924a
Written-by: Reviewer
Tip SHA: e1f924a

Ran on the merge tree (354f3bc5 is this tip's base, so the head tree is the merge tree):

$ node .github/scripts/changelog.mjs check    → changelog: 46 fragment(s) in changelog.d/ are well formed
$ ls changelog.d | grep -E '^(909|910)'      → only the fragment this PR adds; no duplicate for #909 or #910

The name matches changelog.d/<slug>.<kind>.md with security in changelog.mjs's KINDS, the content is one paragraph with no leading - and no heading, and it carries the issue link. The diff is that one file: nothing for arch-docs to ask for, and all 13 checks are green at this tip.

Two body-only notes, no tip change: the body says .changelog.d/ where the path is changelog.d/, and Fixes #909 belongs on #910, which closed it — this one is Refs #909 / Refs #910. I am not holding the merge for that; fix the two lines if the squash message should read accurately.