Authorized security assessments · Zürich
Swiss-sovereign
AI security assessments.
Frontier open-weight security models, run on Swiss GPUs or inside your own network. Your code, logs and findings never leave Switzerland.
- Processing
- Switzerland only
- Third-party APIs
- None
- Findings
- Human-validated
- Pricing
- Fixed after scoping
Process
Four steps. No surprises.
Every engagement follows the same controlled sequence. Nothing is tested before it is authorized in writing, and nothing changes state without a human saying yes.
- 01
Scope and written authorization
We agree targets, time windows and rules of engagement in a short call. You sign a written authorization before any test traffic is sent.
- 02
Isolated deployment
Models run in a Swiss datacenter or on an appliance inside your network. No outbound network access beyond the agreed targets.
- 03
Agentic assessment
AI agents test within a technically enforced scope allowlist. Any state-changing action is paused until a human approves it.
- 04
Human-validated report
Every finding is reproduced and signed off by a human, with evidence, severity and remediation. One retest of fixed findings is included.
Difference
Built for organizations whose data cannot leave the country.
Swiss-only processing
Inference runs on GPUs located in Switzerland or on your own hardware. No US or Chinese model APIs are involved at any point.
Data processing agreement
A DPA aligned with the Swiss FADP (nDSG) and the EU GDPR is part of every engagement, with retention and deletion terms agreed up front.
On-prem option
For regulated environments, the full stack runs as an appliance inside your perimeter, without internet access.
Human sign-off
Automation finds candidates; people confirm them. No finding reaches your report without manual verification.
Fixed prices
You receive a fixed quote after the scoping call. No open-ended day rates.
Days, not weeks
Agentic testing shortens the active test phase. The delivery date is fixed in writing at scoping.
Services
Assessments
Each service ends in the same deliverable: a human-validated report with evidence, risk rating and remediation guidance.
Web app and API assessment
Authentication, authorization, injection, business logic and API abuse, guided by OWASP ASVS and the OWASP API Security Top 10.
Web / APIExternal perimeter
Internet-facing hosts and services, exposed admin interfaces, TLS and DNS configuration, and forgotten assets.
PerimeterInternal network and Active Directory
Assumed-breach testing from inside: segmentation, privilege escalation paths, Kerberos and AD misconfigurations.
Internal / ADCloud configuration review
IAM, network exposure, storage, logging and key management on AWS, Azure, GCP and Swiss providers, with read-only access.
CloudContinuous assessment
Recurring, scoped testing of your external surface and key applications. New findings are validated and reported monthly.
SubscriptionOn-prem appliance
The assessment stack on dedicated hardware inside your network, for environments where nothing may leave the premises.
Appliance
Pricing
Indicative pricing
Indicative starting prices in CHF, excluding VAT. After the scoping call you receive a fixed price for your exact scope.
| Web app and API assessment | fromCHF 3,500 |
|---|---|
| External perimeter | fromCHF 4,500 |
| Cloud configuration review | fromCHF 4,000 |
| Continuous assessment | fromCHF 1,500/ month |
| On-prem appliance | On request |
Fit
Who it is for
Swiss SMEs and SaaS or fintech suppliers who need credible, independent evidence of their security posture.
ISG reporting duty
Operators of critical infrastructure must report cyberattacks to the Federal Office for Cybersecurity (BACS) under the Information Security Act. Knowing your exposure beforehand matters.
FINMA Circular 2023/1
Banks must manage ICT and cyber risks, including regular vulnerability assessments and penetration tests, and pass these expectations on to critical suppliers.
NIS2 supply-chain questionnaires
EU customers in scope of NIS2 pass security requirements down to their suppliers, including Swiss ones.
Customer security reviews
Enterprise procurement asks for a recent third-party penetration test. A current, validated report answers that question directly.
An assessment supports these obligations. It does not by itself establish compliance.
Trust and legal
Rules we work by
Responsible disclosure
Found a vulnerability in our own systems? Please report it. We respond to every good-faith report.
security@icewall.chWritten authorization
Every engagement runs under a signed authorization and agreed rules of engagement.
Agreed scope only
We test only the systems, accounts and time windows you approved. Scope is enforced technically, not just contractually.
Data deletion
Engagement data, including logs and evidence, is deleted at the end of the agreed retention period.
Approval before impact
No exploitation, write or configuration change happens without explicit approval by a human tester.
Start with a scoping call.
We clarify scope, constraints and timing, then send you a fixed quote.
Book a scoping call