fix: suggest_questions bridge-node question restated hub's own name as a target

[cybersecurity-cell]

Summary

• suggest_questions()'s bridge-node template names the bridge node's own community label right after the node's name: f"Why does \{label}` connect `{comm_label}` to {other_labels}?"`.
• An extreme hub (highest betweenness) is frequently isolated by Louvain into its own singleton community, auto-named after that same node. When that happens, comm_label collapses to (approximately) label, and the generated question reads: Why does \useLang()` connect `useLang` to `Settings`, `Chart`?` — the node's own name appears twice, the second time framed as if it were a distinct thing it connects to.
• Found this on a real ~3,150-node project graph: 3 of the top-3 bridge-node questions all hit it, since the highest-degree nodes are exactly the ones most likely to end up alone in their own community.
• Fix: when a normalized (lowercased, non-alphanumeric-stripped) comparison of comm_label and label matches, drop the "connect {comm_label} to" framing and use f"Why does \{label}` bridge {other_labels}?"` instead. Left unchanged when the two labels genuinely differ (verified against a case in the same report where they do).

Test plan

• Added test_suggest_questions_bridge_node_no_self_referential_connect to tests/test_analyze.py, reproducing the bug on a minimal 3-node graph (fails without the fix, passes with it)
• Full pytest tests/test_analyze.py — 50 passed
• Manually re-verified against the real project graph that surfaced this: the three affected questions now read cleanly, and the one genuinely-distinct case is untouched

[graphify-labs]

Graphify reviewed this change.

Looks safe to merge — no coupling regressions and no blocking issues, checked against the code graph (not a self-assessment).

---

Graphify review — findings

This PR modifies the suggest_questions function in graphify/analyze.py to change how bridge-node questions are phrased when a node's own community label matches the node's own label (e.g., a hub isolated into its own singleton community). In that case it now generates a "Why does X bridge ...?" phrasing instead of the "Why does X connect X to ...?" phrasing, using a normalized comparison of the labels. A corresponding test is added to tests/test_analyze.py covering this self-referential scenario.

No blocking issues surfaced. 1 lower-confidence candidate did not survive cross-model review.

Analysis details — impact, health, verification

Impact & health

Graphify review

Impact — 526 functions depend on the 110 functions this change touches.

Health — this change adds coupling hotspots:

• worse: suggest_questions() — 12 callers, 4 callees

Verification — 526 functions in the blast radius were not formally verified this run (proofs are advisory here).

Gate & verification

graphify gate

PASS — objectively clean (no health regressions, tests not run — proofs not run this pass (advisory)). Grounded, not self-assessed.

Advisory (not blocking):

• verification_scope: 365 function(s) in the blast radius were not formally verified this run

· 1 more finding(s) on lines outside this diff (see the check run).