AI-powered scanner vulnerabilities
Summary: PortSwigger Research describes how AI-powered scanners that use LLMs to interpret web content can be manipulated through indirect prompt injection, where malicious instructions embedded in stored content are treated as part of the scanner's action plan. The research notes these scanners often run inside internal networks with authenticated sessions and API access, making them a programmable SSRF vector through mechanisms such as open redirects, URL interpretation discrepancies, and Host header manipulation.
Key Takeaway: A scanner that reads attacker-controlled pages and plans its own next actions is a privileged actor inside your network - one that can be tricked into posting retrieved credentials publicly or reaching internal services an external attacker could never touch.
Read the original at PortSwigger Research ↗
- Topics
- Vulnerabilities & Patching
- Related
- Attackers Exploit UnrealIRCd Flaw That Enables Server-Level Command Injection Cyber Risks · Nov 14, 2026U.S. CISA adds Check Point and UnrealIRCd flaws to its Known Exploited Vulnerabilities catalog Cyber Risks · Nov 14, 2026PostgreSQL Patches 52 Flaws, Including 22-Year-Old Bug Enabling Authenticated RCE Cyber Risks · Nov 13, 2026