Security
Last updated 13 September 2026 · applies to helpmedump.com
What we actually do, rather than a list of words like military-grade. If you find a hole, the last section tells you how to tell us and what we promise in return.
Accounts
- There are no passwords to leak. Sign-in is a one-time link, valid for fifteen minutes, usable once. The link is never stored: only a hash of it is, so a database leak does not hand anybody a working sign-in.
- Sessions are signed, HttpOnly, Secure and SameSite=Lax, and are re-checked against the account on every request so revoking access takes effect immediately rather than in thirty days.
- Every request is authorised on the server against the record being touched. Nothing trusts an identifier sent by a browser.
Data
- Verification documents are never public. They are reachable by operators only, and the check for that is on the server.
- A quote room link is a capability, so those URLs are excluded from search engines and never indexed.
- Analytics strip emails, phone numbers, full postal codes and street addresses before anything is stored.
- Deletion deletes. It does not set a flag and keep the row, and backups roll off within thirty days.
- Nightly encrypted backups, thirty kept, with restores tested rather than assumed.
Platform
- HTTPS everywhere, with HSTS, and certificates renewed automatically.
- A content security policy, frame denial, referrer restriction and a permissions policy, sent on every response.
- Rate limits on sign-in and on anything that sends mail, to make credential and mail abuse expensive.
- Input is validated against a schema at the boundary, and the database enforces the rules that matter rather than relying on application code to remember them.
- Administrative actions are written to an audit log that operators cannot edit.
Reporting a vulnerability
Email security@helpmedump.com. Tell us what you found and how to reproduce it. In return:
- We acknowledge within two working days and tell you what we think.
- We will not threaten you with legal action for a good-faith report.
- We fix genuine issues and tell you when they are fixed.
- We will credit you publicly if you want that, and not if you do not.
Please do not run automated scans against production, access anybody else’s data, or degrade the service for other people while testing.