ci: enforce the 'never surveil' brand guardrail as a copy-word CI check

[nandal]

Background

Our brand guardrail (see CLAUDE.md) is firm: iEye watches over you, never watches you. Copy is always "looking out for / you are seen / never alone" — never "monitor / track / surveil." Today this is enforced by convention only. Per our "enforce structurally, not by convention" principle, let's make it a real CI gate.

The task

Add a small check that scans user-facing copy for a denylist of surveillance words and fails CI on a hit, with an allowlist for legitimate phrases (e.g. "watch over").

• A node or shell script, exposed as npm run lint:copy in landing/, plus a CI step.
• Denylist (seed, refine in the PR): monitor, surveil, surveillance, spy, track you, watch you, camera, eavesdrop.
• Allowlist: watch over, looking out, oversee.
• Scope to copy, not code: start with landing/src/components/HonestyLine.tsx, landing/src/lib/legalCopy.ts, landing/src/pages/*.tsx, and the app's user-facing strings. Don't scan technical identifiers/comments (a "tracking issue" reference must not trip it).

Acceptance criteria

• Adding "we monitor you" to a copy file turns CI red; "we watch over you" stays green.
• npm run lint:copy runs locally and in CI.
• A one-line note in CONTRIBUTING.md pointing to the check.

Why this is a good first issue

Self-contained, needs no app architecture knowledge, and it directly protects a load-bearing guardrail. Pointers: CLAUDE.md, docs/standards/.

[wrokin]

🤖 wrokin triage

Add a CI linting step to enforce brand copy guidelines by scanning user-facing text for prohibited surveillance terms, with an allowlist for approved phrases.

Suggested priority: medium
Labels applied: enhancement, good first issue, javascript
Possible duplicate: no

_{Model: deepseek/deepseek-v4-pro · your key, your model (BYOK)}