Trust & Security

We’re a security company,
and we act like it.

Proving an exploit is real takes source code, credentials, and a live staging environment. This page sets out exactly what Xora does with each one — and what it never does.

What a pentest takes

  • Your source code
  • Staging credentials
  • Reach to your declared target

What it leaves behind

  • Findings and their evidence
  • Execution logs from the pentest

Never your source code.

Our security principles

Four commitments the architecture enforces

Not policies written down and hoped for. Each one is a property of how the system is built, which means it holds even when an agent misbehaves.

We don't store your source code

Whitebox analysis runs in ephemeral memory — read, scanned, discarded. Nothing persists to disk or survives the pentest.

Account data is encrypted at rest

AES-256 across findings, reports, and account records. Credentials carry a further layer — column-level encryption with per-tenant keys, decrypted only inside the sandbox at the moment of use and never written to a log.

Every pentest is its own isolated environment

A fresh, single-use sandbox per pentest — including repeat pentests from the same customer. No shared file system, process, or memory. Destroyed on exit.

Agents can't reach what you haven't authorized

Network egress is locked to your declared target scope at the infrastructure layer — a network guarantee, not a policy an agent could talk its way around.

Architecture & data handling

What a pentest actually touches

A pentest runs in a sandbox with no route to the internet. Its only path out is an egress proxy that permits the hosts you declared and denies everything else. When the pentest ends, the sandbox and everything in it is destroyed.

Your CI pipelinepre-deploy steppentest requestXora control planeRBAC · audit log · scopedispatch + scoped credentialsEPHEMERAL SANDBOX — ONE PER PENTESTprivate subnet · no internet gateway · no shared filesystem, process, or memorySource, in memoryread-only, scoped cloneWhitebox analysisnever persistedExploit agentscredentials decrypted hereOn exit: sandbox, source, and decrypted credentials are destroyed. Nothing is carried to the next pentest.findings onlyFindings + evidenceAES-256 at restkept until you deletethe only route outL7 egress proxydeny by defaultallowYour declared targetstaging only, never productiondenyevery other host — including ours
One pentest, end to end. The sandbox has no internet gateway, so the proxy is not a checkpoint an agent chooses to pass through — it is the only route that exists. The single accent path leaving the sandbox carries findings; source code has no path out at all.

Staging only, never production

Xora connects exclusively to staging and pre-production environments, removing production outage risk by design.

Encryption everywhere

TLS 1.2+ in transit, AES-256 at rest. Credentials use per-tenant column-level encryption and are decrypted only transiently, inside the sandbox.

Tenant isolation by construction

Every pentest gets a dedicated sandbox — including repeat pentests from the same customer — with no shared file system, process, or memory.

Scoped access and audit logging

RBAC governs who can configure scope, view findings, and manage credentials. SSO is available for enterprise. Admin actions and scope changes are logged.

Controls

How each control is actually enforced

The layer a control lives in determines what can bypass it. These are grouped by where enforcement happens, not by what the control is called.

Security controls, how each is enforced, and the layer that enforces it
ControlEnforcementLayer
Encryption in transitTLS 1.2+ on every connection between you, the control plane, and the sandboxAll surfaces
Encryption at restAES-256 across the findings database and evidence storageStorage
Credential storageColumn-level encryption with per-tenant keys; ciphertext is all the control plane holdsControl plane
Credential useDecrypted inside the sandbox at the moment of use, never logged, destroyed with the sandboxPentest sandbox
Source codeCloned into memory for the duration of the pentest; never written to disk, never persistedPentest sandbox
Tenant isolationA dedicated single-use sandbox per pentest — no shared file system, process, or memoryPentest sandbox
Network egressDeny by default. Private subnet with no internet gateway; all traffic forced through an allow-listed L7 proxyNetwork
Environment scopeStaging and pre-production targets only — production is out of scope by designProduct
Model provider retentionZero-data-retention terms with every inference provider — no storage after the request, no training on customer dataVendor agreements
Access controlRBAC over scope configuration, findings visibility, and credential managementControl plane
Single sign-onAvailable for enterprise accounts via your identity providerControl plane
Audit loggingAdmin actions and scope changes recorded and retainedControl plane
Subprocessors

Who else is in the path

The third parties that process customer data on Xora’s behalf, and what each one sees. We notify customers before adding a subprocessor that changes this list.

Subprocessors, their purpose, and the data each one processes
SubprocessorPurposeData processed
Amazon Web ServicesSandbox compute, secrets management, evidence storagePentest data in memory; encrypted evidence at rest
NeonManaged Postgres for findings, reports, and account recordsFindings, organisation metadata, encrypted credentials
VercelHosting for the dashboard and marketing siteRequest metadata
WorkOSAuthentication, SSO, and directory syncUser identity and session data
AnthropicModel inference for analysis and exploit agentsCode and response excerpts during a pentest
OpenAIModel inference for analysis and exploit agentsCode and response excerpts during a pentest
BasetenModel inference for analysis and exploit agentsCode and response excerpts during a pentest
LangfuseCost and latency observability for model callsPentest metadata only — token counts, model IDs, timings. Prompt and response content never leaves our VPC.
Merge.devTicketing integrations (Jira, Linear)Finding summaries you choose to sync
ResendTransactional and digest emailEmail address, finding summaries
GoogleMarketing site analyticsVisitor IP and page activity on getxora.ai. No customer pentest data.

Every model provider we use runs under zero-data-retention terms. Prompts and responses are not stored once the request completes, and are never used to train models — yours or anyone else’s.

Compliance & assurance

Where we stand today

Compliance frameworks and Xora's current status against each
FrameworkStatusNotes
SOC 2 Type IINot yet certifiedReadiness work is underway and we are selecting an audit partner. We will publish the observation window start date once it is set.
ISO 27001RoadmapPlanned following SOC 2 Type II completion.
GDPRAlignedThe staging-only model limits personal data exposure. Data processing terms are agreed per contract.

Ahead of formal attestation, our architecture — ephemeral execution, per-tenant encryption, staging-only scope, destroyed-on-exit sandboxes — is designed to satisfy the substance of the controls SOC 2 attests to.

Incident response & retention

What happens when something goes wrong

Responsible disclosure

Report a suspected issue to security@getxora.ai. We acknowledge within two business days.

Incident notification

Confirmed incidents affecting customer data are disclosed without undue delay, per your agreement with us.

Retention

Source code and sandbox state are never persisted beyond the pentest. Findings, evidence, and the execution log of the pentest remain in your account until you request deletion.

Xora is responsible for

  • Securing the platform and every pentest environment
  • Isolating and destroying each pentest sandbox
  • Enforcing egress scope at the network layer
  • Encrypting credentials, findings, and evidence
  • Notifying you of confirmed incidents

You are responsible for

  • Scoping engagements accurately
  • Confirming targets are staging, not production
  • Managing platform access and roles
  • Rotating the credentials you supply
  • Remediating the findings we return
Security FAQ

The questions reviewers ask

What happens to our source code after a pentest?

Nothing, because nothing was kept. Source is cloned into the sandbox's memory, analysed, and discarded when the sandbox is destroyed. It is never written to disk and never leaves the sandbox — what leaves the sandbox is the findings, their evidence, and the execution log of what the agents did.

Can an agent reach a system we didn't authorise?

No. The sandbox runs in a private subnet with no internet gateway, so there is no route to the internet for an agent to find. Its only egress path is a proxy that permits the hosts in your declared scope and denies everything else. This is enforced below the agent, so a confused or compromised agent cannot reason its way past it.

Do you ever test production?

No. Xora connects only to staging and pre-production environments. That is a design decision, not a setting — it removes production outage risk entirely, and it is why the platform can attack aggressively enough to prove an exploit rather than just flag a pattern.

Do you use our code or findings to train models?

No, and neither do our model providers. Your code and findings are used only to carry out your pentests and produce your reports. Every inference provider we use runs under zero-data-retention terms, so prompts and responses are not stored once the request completes and are never used as training data.

Who at Xora can see our findings?

Access is governed by RBAC within your organisation. Xora personnel access customer findings only when you request support, and those actions are logged.

What if an exploit breaks our staging environment?

Staging is where that risk belongs, which is the point of the staging-only model. Pentests are scoped to the targets you declare, and every action an agent took is in the pentest's evidence trail, so anything that broke can be traced to the request that caused it.

Can we see the SOC 2 report?

Not yet — we are not certified, and the observation window a Type II report covers has not started. We would rather tell you that than imply otherwise. The white paper and a signed NDA for deeper technical review are available in the meantime, and we will say so here the moment that changes.

Send this to your security reviewer

The white paper covers everything on this page in a document your reviewers can file — architecture, controls, compliance status, and incident response.

Get our security whitepaper

Need an NDA for deeper technical review? security@getxora.ai