Skip to content
NewA calmer, faster ZeroHour

The Blind Spots of Agentic Pentesting and How to Close Them

BleepingComputerby Sponsored by Picus Security ()6 min read3 · 1 read
AI summary · grok-4.7-buildSummaries by Vani AI

Picus says agentic pentesting misses unreachable assets and flaws that lack safe exploits.

A Picus Security research engineer argues that agentic penetration testing is creative but inconsistent, token-expensive, and limited to assets it can safely reach. Omdia found organizations pentest about 32% of their attack surface, while safety and segmentation leave critical or isolated systems untested. In one customer environment, scanners found nginx-ui authentication bypass CVE-2026-33032 on 480 instances, but agentic testing could assess only 120 and confirmed 80 exploitable. VulnCheck reported that only about a quarter of 2025 CVEs had proof-of-concept code, so Picus advocates control-based exploitability validation alongside deterministic and agentic testing.

Key facts
  • Agentic tests can only reach assets that are safe and network-accessible.
  • In one case, 360 of 480 nginx-ui instances were out of scope.
  • VulnCheck says about a quarter of 2025 CVEs had public exploit details.
  • Picus recommends pairing frequent deterministic tests with periodic agentic runs.
Analyze this story with Vani AIImpact, who is affected and what defenders should do, in a new chat

Vulnerabilities mentioned

All
CVEVulnerabilityCVSSEPSSSignalsPublished
CVE-2026-33032
Unauthenticated MCP Endpoint Access in Nginx UI Enables Full Nginx Server Takeover
9.835%PoC

From the original article

6 min read at bleepingcomputer.com

AI-powered penetration testing can uncover complex attack paths, but safety restrictions and network limitations leave many assets and vulnerabilities untested. Picus Security explains how combining agentic pentesting with exposure validation and breach simulation can help close these gaps. [...]

BleepingComputer

By Sponsored by Picus Security · bleepingcomputer.com

Read the full article at BleepingComputer

A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.