security
Security and responsible disclosure
What we store, how the controls work, what has and has not been independently checked, and how to report a vulnerability. Last updated: .
What an account holds
Your e-mail address is the account identifier. Operational records linked to it include sign-in times and role, plans and purchases, Academy progress, messages and content you send us, security events (with IP address and a truncated browser identifier), and — if you use them — scan targets and reports, SOC events, assets, incidents and workspace settings. The full list, with retention, is in the Privacy Policy. What we store
Sign-in and sessions
Sign-in uses a six-digit code sent by e-mail. Codes expire after 5 minutes, work once, are stored only as a keyed hash, and are rate-limited per address, per network and per account. E-mail codes are a single factor, not multi-factor authentication; passkey (WebAuthn) support is planned. Sessions end after 30 minutes idle or 12 hours after sign-in, and you can see and sign out your other sessions from your account page.
Encryption
All traffic uses TLS, with HSTS (one year, including subdomains). Data at rest is stored in the hosting provider's database and file system; the application does not add its own encryption layer to that data. Secrets we issue (sign-in codes, site-verification tokens, agent and API keys, team-invite tokens) are stored as hashes, not in readable form. Quantum Vault encrypts text with AES-256-GCM on the server in memory and stores nothing; it is not end-to-end encryption.
Application safeguards
Session cookie: HttpOnly, Secure over HTTPS, SameSite=Lax. CSRF tokens on state-changing forms. Content-Security-Policy with script-src limited to this site for visitors who have not opted in to analytics or advertising (with that consent, Google origins are added and inline scripts are allowed; inline styles are allowed for everyone). Outbound fetches of customer URLs block private, loopback, link-local, metadata and other reserved addresses. AI-generated code is never executed on the server.
Website scanning
Scans are passive: they read public responses and DNS records and do not attempt exploitation. One-off scans require you to confirm you are authorised. Recurring monitoring additionally requires proof of control of the domain (DNS TXT record or a file on the site), which is re-checked periodically. Scans are rate-limited per account, per organisation and globally.
Assurance — what has been checked
Controls are covered by an automated test suite and by internal code reviews (self-assessed). NeoShield has not had an independent penetration test and holds no security certification (such as SOC 2 or ISO 27001). We say so plainly rather than imply otherwise.
Who operates it
NeoShield is operated from Japan. There is no on-call rota and no SLA. Backups, monitoring and what happens outside support hours are described in the Trust Center. Operational continuity
Responsible disclosure policy
We welcome reports from security researchers. The full policy — scope, safe harbour, what qualifies (including rate-limit, authentication-abuse and AI-cost issues) and what is out of bounds — is on the responsible disclosure page.
Report privately to security@neoshieldsecurity.com. Machine-readable details are at /.well-known/security.txt.
Service providers
Hosting: Hostinger. Payments: Stripe (card details are entered on Stripe-hosted pages and never reach our servers). AI features: Anthropic. Analytics and advertising (only with your consent): Google, and PostHog if enabled. Purposes and processing locations are listed in the Privacy Policy. Processors