fix(uxcore): address PR #181 review comments

[manager]

Follow-up to #181: addresses the review bot's inline comments that were merged past.

• OffsecBiasView: conditional classNames now use the cn() helper per AGENTS.md (was raw template strings).
• rating.ts: geo payload typed as TUserGeo instead of an explicit any.
• OngoingLayout.module.scss: submit-error color is the shared error red as a named variable, not a fresh hardcoded literal.

The fourth review note (sign-off for the OffSec bias-schema additions) needs no code change: the content and its dev-only gating were explicitly approved by the maintainer.

🤖 Generated with Claude Code

[github-actions]

Claude finished @manager's task in 1m 53s —— View job

---

Claude's PR review

• Read trigger request and repo conventions
• Review src/uxcore/api/rating.ts
• Review src/uxcore/components/OffsecBiasView/OffsecBiasView.tsx
• Review src/uxcore/layouts/OngoingLayout/OngoingLayout.module.scss
• Post overall assessment

Summary

Small, well-scoped follow-up that does exactly what the description says. No bugs, no convention violations, nothing blocking.

rating.ts — any replaced with a proper TUserGeo shape; the four destructured fields (country, region, city, ip) map exactly to the type. Minor, non-blocking nit: the type TUserGeo alias is declared inside the function body (rating.ts:8) rather than at module scope — works fine here since it's single-use and not shared (so no .types.ts file is warranted per AGENTS.md), but hoisting it above rateRequest would read slightly more conventionally. Not worth a revision on its own.

OffsecBiasView.tsx — all three conditional/multi-class className strings now use cn() as required by AGENTS.md. Import order is correct: @uxcore/data/biasOffsec and classnames both fall into the "third-party/catch-all @?\w" ESLint group (since @uxcore has no dedicated alias bucket in the flat config), and within that group @uxcore... sorts before classnames alphabetically — so simple-import-sort is satisfied as-is.

OngoingLayout.module.scss — hardcoded #fb1717 promoted to a $error-red SCSS variable with a comment explaining it matches DecisionTable.module.scss's error color. Note this is a conceptual pairing, not a shared source of truth — the DecisionTable file (lines 151, 292) still hardcodes #fb1717 independently, so the two will silently drift if either changes. That's consistent with this repo's current SCSS conventions (no shared token file yet) and matches what the original review comment asked for, so not blocking — just flagging for awareness if a future pass wants a real shared token.

Item 4 (bias-schema sign-off) — correctly left as no-op per the PR description; nothing to verify code-side.

No inline comments needed — nothing here rises above a nitpick.