Approved under Anthropic’s Cyber Verification Program
Security evaluations assisted by frontier AI
Automated scanning finds known weaknesses. A skilled tester finds how they combine. Conscient Systems now pairs hands-on testing with Anthropic’s Claude models, which lets us cover more ground in the same testing window, reason about your application the way an attacker would, and write findings you can act on.
Conscient Systems is an approved participant in Anthropic’s Cyber Verification Program for basic penetration testing, red teaming and bug bounty use cases. Every engagement is planned and supervised by us, under a written authorization that names the exact scope.
What we test
Basic penetration testing
Manual and AI-assisted security testing of the in-scope targets for exploitable vulnerabilities, including but not limited to the OWASP Top 10, authentication and session handling, access control, injection, misconfiguration and exposed services, with safe, minimal exploitation to demonstrate impact.
Red teaming
Goal-oriented adversary simulation against the in-scope targets, in which the Provider chains vulnerabilities and misconfigurations to reach objectives agreed with the Client, emulating the tactics of a realistic external attacker. Social engineering, phishing and physical intrusion are excluded from this authorization.
Bug bounty style research
Open-ended vulnerability research across the in-scope targets for the duration of the testing window, in the manner of a private bug bounty programme, with each confirmed finding reported to the Client as it is validated.
How it works
- 1
Describe the scope
Sign in to VulnScan and list exactly what may be tested: URLs, domains, IP ranges, the kind of testing, the testing window and an emergency contact.
- 2
Sign the authorization
VulnScan generates a penetration testing authorization and rules of engagement contract in English. You sign it electronically, then Conscient Systems countersigns.
- 3
Agree the details
We contact you to agree on price, timing and practicalities. Testing starts only inside the signed scope and window.
- 4
Get the findings
Critical issues are reported as soon as they are confirmed. You receive a written report with severity ratings and remediation guidance.
Built on consent, scope and supervision
- Nothing is tested without an authorization signed by you and by Conscient Systems.
- Only the listed targets are in scope; everything else is out of scope by default.
- No denial-of-service, social engineering or physical intrusion.
- You can stop the testing at any time, with immediate effect.
- AI-assisted activity is directed and reviewed by Conscient Systems, bound by the same rules as manual testing.
- Findings stay confidential; raw testing data is deleted within 90 days of the final report.
Anthropic and Claude are trademarks of Anthropic, PBC. Anthropic is not a party to our engagements and does not perform, endorse or guarantee the testing.