Exactly what we keep, and what we don't

Who you are

Your work email

Used once to prove you work in IT, then deleted after an admin approves or declines you. It's never shown to anyone.

Your LinkedIn link or vouch

Used only for that review, then deleted with the decision. The same goes for the LinkedIn link used to verify a real name.

How you sign in

A personal email address and a one-time code. No passwords. You stay signed in on each device until you sign out.

Your profile

A handle, plus your role, industry and company size, which are shown as broad groups unless you choose more detail.

Your real name

Only if you choose to post as yourself, only after we verify it, and only on posts where you pick that option. You can turn it off any time; past posts made as yourself then show as "Verified member", never your handle.

What you run

Your stack

The products you run tune your own feed and alerts. They're never shown to anyone, and they're encrypted in the database with a key kept outside it, so a stolen database or backup alone doesn't reveal them.

Alert emails and channel messages

By default they say only how many alerts are waiting, never product or vulnerability names. Teams and Slack webhook links are stored encrypted too.

What you share

Reports and ratings

Patch, upgrade, renewal and outage reports and vendor ratings store no member ID. A one-way code lets you change your own answer and stops double counting. The code can only be linked back to you with the app's secret key, which is kept outside the database, so a stolen database alone doesn't reveal who sent what.

Anonymous posts

Store no author at all. Handle posts and real-name posts are shown separately and count helpful votes separately, so other members can't match a handle to a name.

Minimums and batches

Totals only appear once enough members have reported, and scorecards and pulses update in batches every few hours, so one report can't be picked out. Patch and outage reports show right away because speed matters, with dates rounded to the day and small counts hidden.

Notes you write

Emails, links and phone numbers are stripped from vendor notes. Leave out names and hostnames; admins can hide notes that identify someone.

What our systems record

Your browser never talks to our database

All data goes through our server, so the database never sees your IP address.

Hosting logs

Our hosting provider keeps standard short-lived request logs (IP address, time, page) for debugging and abuse. Our own code doesn't log what you report, rate or post.

Providers

The app runs on Vercel, data lives in Supabase, sign-in codes come from Supabase Auth and alert emails are sent through Resend. AI tagging of public news runs on Anthropic's Claude and never receives member data.

If something goes wrong

A database breach

Would expose handles, broad profiles, posts and totals. Verified names are stored apart from members with no member ID, so a breach couldn't tie a name to a handle or to anything posted under one. It wouldn't expose stacks or webhook links (encrypted), work emails (deleted), or who wrote anonymous posts or sent any report (no member ID is stored with them).

A legal demand

We can only hand over what we hold, and the design keeps that small. The database doesn't store which member sent a report, rating or anonymous post. Recomputing that link would take the app's secret key, which is never stored in the database.

Still to come

Identity data in its own separate database, passkey sign-in, an independent penetration test with a public summary, self-serve data export and account deletion.