Sovereign · self-hostable · authorized-scope only

The self-hostable pen-tester that proves every finding — and only touches what you authorize.

SENTRY is an autonomous AI penetration tester that proves vulnerabilities with working exploits, not scanner noise. It runs entirely in your environment, and it physically cannot test a target outside a scope you have declared, attested and had approved. UK/EU-hostable, no US CLOUD Act exposure.

Attested-scope onlyIn-code allowlist + kill-switchValidated exploitation + PoCSelf-hosted / air-gappedSecure-comms crypto tests

Built for defenders — safe by construction.

An autonomous offensive tool is only responsible if it cannot be pointed at a target you have not proven you may test. SENTRY enforces that in code, not in a prompt.

Authorized scope only

Every network-touching probe passes an in-code allowlist first. A target is permitted only if it belongs to an engagement you have declared, attested authorization for, and a human has approved. No attestation, no scan.

Validated, not noisy

Every finding is reproduced with a second confirming request before it is filed, and re-checked adversarially — cutting the 40–70% false positives raw scanners produce.

Sovereign by default

Runs in your environment — on-prem, air-gapped, UK/EU. Your targets, findings and traffic never leave the box. Swap the reasoning brain for local inference and nothing phones home.

How access works

SENTRY is not anonymous self-serve. You cannot point it at an arbitrary URL. Access is granted per-engagement, the way every responsible offensive-security platform works.

Declare your scope + attest authorization. You list the exact hosts you own or are contracted to test, and attest — on the record, with your name, org and legal basis — that you are authorized. Start here.
A Pharoah admin reviews and approves. Nothing runs on submission. A human approves the declared scope (time-boxed) before the engine can touch it. Signup never triggers a scan.
The engine runs — inside your approved scope only. The approved hosts populate the engine's allowlist. Any target outside it is refused, fail-closed, with a kill-switch that stops everything instantly.
Every finding is proven and logged. Reproduced with a PoC, written to your report with severity, evidence and a fix. Every scan attempt — permitted or refused — is recorded for audit.

Real, reproduced findings

On an authorized run against a staging secure-messaging relay, SENTRY confirmed these with working PoCs — 0 false-positive criticals, every auth gate and the crypto core independently verified sound. Pointed at a separate deliberately-vulnerable API, the same engine reproduced a real password-leaking endpoint and missing-control findings — each confirmed by an independent second request before filing.

Inbox-metadata BOLA

Unauthenticated read of a victim's inbox metadata — sender graph, timing, volume. Content stayed encrypted; the metadata didn't.

Excessive data exposure

An endpoint returned populated password fields in its response body. Reproduced, then reported with the exact fix.

Crypto red-team

Offline P-256 ECIES / AES-GCM / ECDSA harness — AEAD tamper, invalid-curve, cross-recipient and forward-secrecy properties, the QA a scanner cannot do.

What it actually is

A profile-driven REST-API vuln engine — missing security headers, broken authorization, excessive-data-exposure, boolean/error-based SQL injection, rate-limiting — target-agnostic, every VULN reproduced.
An offline crypto integrity harness for secure-comms — P-256 ECIES/AES-GCM/ECDSA vectors that verify AEAD integrity, invalid-curve rejection and forward-secrecy properties.
An in-code allowlist + kill-switch — the load-bearing safety guard. Scope is enforced in code, not in a prompt, and extended only by approved, attested engagements.
Optional autonomous mode — the same hard-scoped tools driven by a multi-model reasoning brain, for a full recon → exploit → verify → report kill-chain, still inside the approved scope.

Pricing

The engine costs £0 to run, so continuous per-commit testing doesn't burn a per-test budget. Access is sold per authorized engagement — purchase does not grant instant, unattested scanning.

Team · self-hosted / CI
£299 / month
  • Full pen-test engine — deterministic suite + optional autonomous mode
  • Self-host on your own infra (Docker / air-gapped)
  • Validated exploitation + PoCs
  • Crypto + secure-comms test suite
  • Per-engagement authorized scope + audit log
  • Email support
Enterprise & Government
Custom
  • Everything in Team, at estate scale
  • Sovereign / on-prem deployment + support
  • Fully-offline local-inference option
  • External-audit bundle (Cure53 / NCC pathway)
  • Procurement-ready assurance pack
  • Priority support + SLAs
⛔ Authorized testing only. SENTRY enforces an in-code host allowlist and a kill-switch, and refuses any target outside a scope you have declared, attested and had approved. There is no anonymous "point it at any URL" mode. You are responsible for operating it lawfully and only against systems you own or are contracted to test — see our acceptable-use terms.

FAQ

Can I just paste a URL and scan it?

No — by design. You declare a scope and attest authorization, a human approves it, and only then can the engine touch those exact hosts. This is the same model responsible platforms (Cobalt, HackerOne, Intruder) use, and it is why SENTRY is safe to run autonomously.

Does anything leave our environment?

No. The engine, the target and the findings run locally. The only optional outbound call is to the reasoning brain — point that at a local model for a fully offline, air-gapped deployment with zero external calls.

How is this different from a vulnerability scanner?

Scanners flag potential issues and produce 40–70% false positives. SENTRY reproduces each finding with a working exploit and has an independent model try to refute it before reporting — so what you get is validated and actionable.

What stops it going out of scope?

An in-code allowlist (not a prompt instruction), fail-closed by default, plus a kill-switch. A target is permitted only if it is an approved, attested engagement scope. Every attempt, permitted or refused, is logged.