AI agents execute
Eight specialized AI agents test your perimeter following a professional pentest methodology.
AI-Driven Pentest · Human in the Loop
AI agents hack. Human pentesters confirm.
Eight specialized AI agents test your perimeter following a professional pentest methodology. Then a human pentester personally validates every vulnerability before it reaches your report, delivered in Portuguese, English or Spanish.
- Every finding validated by a human pentester
- Contracted, authorized and under NDA
- Report in PT · EN · ES
- 8specialized AI agents
- 100%of findings validated by a human pentester
- 0unvalidated findings in your report
- 3report languages
Human in the Loop
01
02
03
04
AI proposes. A human pentester approves.
AI agents execute the engagement and propose findings, but every vulnerability is personally validated by a human pentester before it reaches your report. They reproduce the exploit, confirm the evidence, discard false positives and set the final severity.
Evidence collected
Every potential vulnerability is documented with clear evidence and a reproduction path.
A pentester validates
A human pentester reproduces each finding, confirms the evidence, discards false positives and sets the final severity.
Validated report delivered
Only confirmed vulnerabilities reach the corporate report, in Portuguese, English or Spanish.
AI-only scanners
- Thousands of unvalidated findings
- False positives dumped on your team
- No severity judgement, just raw CVSS noise
- Nobody accountable for the result
NFGuard + Human in the Loop
- Every vulnerability reproduced by a human
- False positives removed before delivery
- Severity set by a pentester, not a script
- A named professional stands behind the report
8 Specialized AI Agents
Eight AI specialists covering your entire attack surface
Each agent focuses on one area of risk, and every result goes to a human pentester before it reaches you.
-
External Attack Surface
We map everything you expose to the internet: domains, services and certificates, including what your team may have forgotten.
-
Web Presence
A complete picture of your websites, portals and the technologies behind them.
-
Web Applications & APIs
Tested for the flaws attackers exploit most: injection, broken authentication and data exposure.
-
Known Vulnerabilities
Your systems checked against publicly known vulnerabilities, focused on what is actually exploitable.
-
Exposed Credentials
Leaked passwords, keys and tokens found before someone else finds them.
-
Software Supply Chain
Risks hidden in the components, containers and dependencies your products rely on.
-
Threat Intelligence
Every finding in context: is it being actively exploited, and how likely is an attack?
-
Reporting
A clear corporate report, written for executives and technical teams alike.
Professional Deliverable
What your team receives
A corporate intrusion-test report on the DolutechAI template.
- Cover and engagement identification
- Executive summary with global risk rating
- Scope and methodology
- Criticality view with CVSS scoring
- Architecture overview
- Detailed findings (VULN-01…N): description, evidence, impact, remediation
- CVEs with CVSS/EPSS context
- P0 to P3 remediation roadmap
- Tests performed without findings
- Strategic recommendations
- Conclusion
Narrative in 3 languages
Portuguese, English or Spanish. The report is written in the language you choose.
Ready to share
Delivered in PDF, DOCX and HTML, ready for your board, your team and your auditors.
Signed off by a pentester
Every vulnerability in the report was reproduced and confirmed by a human pentester. No unvalidated scanner output.
Engagement & Compliance
A pentest you can audit
Every engagement is contracted, authorized and confidential, from scoping to the final report.
-
Written authorization & defined scope
We only test what you authorize in writing: domains, applications and IP ranges agreed before testing begins.
-
Compliance contract & NDA
Every engagement is governed by a compliance agreement and an operator NDA.
-
Operated by DolutechAI specialists
AI agents execute; our pentesters run, supervise and validate the engagement end to end.
-
Evidence & audit trail
Documented evidence and a reproduction path for every finding, so your team and auditors can verify it.
-
Nothing to install
Our team runs the entire engagement. No agents or software deployed in your environment.
-
Security by design
Isolated, hardened operation with strict controls from the first test to the final report.
Frequently Asked Questions
NFGuard FAQ
What is NFGuard?
NFGuard is DolutechAI's AI-driven penetration-testing service: eight specialized AI agents test your perimeter following a professional pentest methodology, and every vulnerability is personally validated by a human pentester before delivery.
Does a human really validate every vulnerability?
Yes. Human in the Loop is the core of the product: a pentester personally reproduces each candidate finding, confirms the evidence, discards false positives and sets the final severity. Nothing unvalidated reaches your report.
Is NFGuard fully automated? Who operates it?
Testing is executed by AI agents, but every engagement is operated by DolutechAI specialists under contract, and results are always reviewed by a human pentester.
What does the report include and in which languages?
A corporate intrusion-test report: executive summary with global risk, scope, methodology, detailed findings with evidence and remediation, CVEs, a P0 to P3 remediation roadmap and conclusions. Delivered in Portuguese, English or Spanish, in PDF, DOCX and HTML.
How does an engagement start?
Tell us your scope: domains, applications or IP ranges. We formalize written authorization, the compliance agreement and NDA; then the AI agents execute the engagement and our pentester validates every finding before the report is delivered.
Is it legal? Do I need authorization?
NFGuard only operates under contract with written authorization, governed by a compliance agreement and operator NDA. Testing third-party systems without authorization is illegal, and it is not what we do.
Put eight AI agents and a human pentester on your perimeter.
Tell us about your scope and we'll design the engagement.