We ask customers to trust us with source code and live systems. Here is what that trust rests on.
Atlas parses repositories to an AST. It does not install dependencies, run build scripts, or evaluate configuration. A hostile repository cannot execute code on our infrastructure.
Repositories are cloned into an isolated, run-to-completion task and the workspace is discarded when it ends. Code is never written to long-term storage.
Ares requires both an approved source repository and DNS-verified host before testing starts. Scope and address checks block private, loopback, link-local, and cloud-metadata ranges. Because testing is intentionally aggressive, the product directs customers to disposable staging environments, never production.
Every organisation-owned table has row-level security, so a query can only ever return rows for organisations the caller belongs to. This is enforced by Postgres, not application code.
Each service holds only the secrets it needs, stored in AWS Secrets Manager and injected at task start. Engine tasks hold no AWS permissions at all.
Evidence captured with a finding passes through redaction before storage, so tokens and credentials observed during a scan are not persisted verbatim.
If you believe you have found a security issue in Sentesting, email security@sentesting.space. Include enough detail to reproduce it.
We do not currently run a paid bug bounty.
We would rather be accurate than impressive. Sentesting is a young product and the following are not yet true:
Our reports are built to support your SOC 2 or ISO evidence gathering. That is different from us being certified, and we will not imply otherwise.