The Blind Spots of Agentic Pentesting and How to Close Them
Agentic pentesting is a new approach to offensive testing, but it has blind spots. This piece examines where they are and how to close them.
Intelligence analysis by Qwen 2.5 (3B)
Agentic pentesting is a new offensive testing method, but it has limitations. This article discusses the blind spots and how to address them.
Agentic pentesting is like when you try to find a hidden toy. Sometimes you can reach it, and sometimes you can't. This article talks about how to find the toys you can't reach and how to make sure you find all the toys, even the ones you can't see yet.
Analysis
The Blind Spots of Agentic Pentesting
Untestable Assets and Exposures
Agentic pentesting can only test what it can safely reach. This limits its coverage, especially in segmented or air-gapped environments. Omdia research found that organizations typically pentest only 32% of their attack surface. Safety restrictions and reachability issues mean that many assets and vulnerabilities remain untested. For example, a customer environment flagged 480 instances of a CVE, but agentic pentesting could only test 120 of them due to network path limitations. This highlights the need for additional testing methods like exploitability validation.
Untested Exploits
Even on assets an agentic pentesting tool can reach, many vulnerabilities can't be tested because there's no exploit to run. Only about a quarter of 2025 CVEs had proof-of-concept (PoC) code or exploit details. This means that some teams may need to rely on LLMs to generate exploits, but doing so carries risks such as potential service crashes or data corruption. Therefore, it's important to have a comprehensive testing strategy that includes both deterministic and agentic pentesting methods.
Closing the Gaps
To address the blind spots of agentic pentesting, security teams should run deterministic tests to cover predictable coverage and agentic tests at set intervals. This allows for consistent testing and reliable retesting of fixes. Additionally, exploitability validation tools like Picus Exposure Validation can help test vulnerabilities that don't have public exploits. This ensures that security teams can identify and mitigate risks effectively.
Conclusion
Agentic pentesting is valuable, but it has limitations. By understanding these limitations and implementing a comprehensive testing strategy, security teams can ensure that they are adequately covering their attack surface and mitigating risks.
Key points
- Agentic pentesting can only test what it can safely reach, leaving some assets and vulnerabilities untested.
- Exploitability validation tools can help test vulnerabilities without public exploits.
- Security teams should run both deterministic and agentic pentesting to ensure comprehensive coverage.
By combining deterministic and agentic pentesting, security teams can ensure they cover all their bases and find all the hidden toys, even the ones they can't reach yet.
If security teams don't use both types of pentesting, they might miss some hidden toys, which could lead to security issues.
