AI Job CounterFree access

Layered security

Bots blocked. Detail gated. Changes require MFA.

The security design combines independent controls instead of relying on one login screen. All activation flags remain off until the required operator, provider, legal and security checkpoints are complete.

01 · Bot and abuseTurnstile before registration or login

Tokens must be checked server-side for hostname and action. Invalid, expired or replayed tokens fail closed.

Prepared · not activated
02 · Detail accessServer-side entitlement on every request

A verified session, completed necessary survey and active access status are all required. Client-side locks never grant data access.

Enforced fail-closed
03 · Data extractionRate limits and non-indexable APIs

Protected responses are no-store and noindex. Durable server-side rate limiting is required before any real detail delivery.

No public detail data
04 · Operator controlTOTP / Google Authenticator at AAL2

Administrative mutations require an operator role, a fresh TOTP step-up and a single-use action nonce.

5-minute maximum proof age
05 · Site integrityNo public editing endpoint

The public application exposes no route that can edit source code, DNS or deployment settings. Provider and repository accounts require their own MFA.

Separate production checkpoint
06 · Audit and recoveryEvery change is attributable and reversible

Source, number, method, approval and revision records are versioned. Production policy cannot be self-approved.

Human review required
07 · Organisation evidenceLogos stay quarantined until reviewed

Raster-only limits, malware scanning, metadata removal, official-domain reference matching and a human decision prevent an uploaded image from becoming public automatically.

Prepared · storage and email disabled
Honest boundary: no website can guarantee that publicly visible text is never copied. The protected datasets therefore never enter public HTML or unauthenticated API responses.