tools/pr/digest.py:518's emit_issues_bd calls bd.list_issues() with no status/include_closed filter, then applies if i.get("status") != "closed" — the same "only literal closed is terminal" pattern fixed in nf-kccf/nf-bhvf (PR #1202), but not swept here.
Bd.list_issues's own docstring confirms real bd list only excludes literal closed by default, so a deferred bd task passes straight through and prints under BD_ISSUES_HEADER indistinguishable from genuinely open backlog: [UNCOVERED], no ownership line, no signal it was explicitly declined.
Not hypothetical — nf-tvc.9 (design 011 T9) is deferred right now and would show up there on the next digest.py run.
Fix: treat deferred as terminal in emit_issues_bd's filter too, alongside closed. Also worth a grep sweep for other bare == "closed" / != "closed" checks against bd status elsewhere in the codebase.
Surfaced by ninefold-reviewer's review of PR #1202 (https://github.com/cfsh/ninefold/pull/1202) — filed unclaimed per that review's own note that it had no bd access to file it directly.
None.
No comments.