Trust
A security product should hold itself to a higher bar.
This is how Sentinel treats your code and everything it learns about it.
Your code
- The scanner reads a checkout and mounts it read-only into each engine. It never edits, commits or runs your code.
- Every engine runs from an exact, pinned version in its own container, removed when it finishes.
Secrets
- Secret scanning runs with redaction on. A finding records the file, rule and commit, never the value.
- Uploads are checked against a fixed set of fields before anything is stored.
Findings
- Stored in the EU (AWS eu-west-1, Ireland).
- Every record belongs to one organisation, and every request is checked against your membership of it.
- Scanners upload with a per-organisation token. Only its hash is stored, and admins can revoke it at any time.
Sign-in
- A sign-in pool dedicated to Sentinel, with multi-factor authentication.
- No open sign-up: your administrator adds people, and a seat is bound to the first identity that uses it.
- Short sessions, in a cookie scripts cannot read and that is sent to this site only.
Accountability
- Triage decisions and admin actions are append-only: superseded, never edited or deleted.
- The audit log records who did what and when, with no IP addresses and no secret values.
Our own standard
- Sentinel scans its own code. Its first scan found a critical advisory in an authentication library we were about to use, so we chose another.
- Every dependency is pinned to an exact version.
Found a vulnerability?
Use the contact form and say it is a security report. We reply to security reports first.
Send a security report