BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//pretalx.com//bsidesluxembourg-2026//talk//UGKRML
BEGIN:VTIMEZONE
TZID:CET
BEGIN:STANDARD
DTSTART:20001029T040000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10
TZNAME:CET
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:20000326T030000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=3
TZNAME:CEST
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
END:DAYLIGHT
END:VTIMEZONE
BEGIN:VEVENT
UID:pretalx-bsidesluxembourg-2026-UGKRML@pretalx.com
DTSTART;TZID=CET:20260508T112000
DTEND;TZID=CET:20260508T120000
DESCRIPTION:AI agents are different from regular LLM apps — they plan ste
 ps\, call tools\, and chase goals across multiple interactions. This added
  complexity introduces new kinds of security risks that aren’t widely un
 derstood yet.\n\nIn this talk\, I’ll walk through demos of vulnerabiliti
 es from the OWASP Agentic AI Threats. These include goal hijacking\, align
 ment faking\, orchestration misuse\, and time-based attacks that exploit h
 ow agents behave over multiple steps or sessions. I’ll show how attacker
 s can trick agents into following the wrong goals\, leaking data\, or usin
 g tools in unsafe ways — all through practical examples.
DTSTAMP:20260412T024938Z
LOCATION:IFEN room 2\, Workshops and AI Security Village  (Building D)
SUMMARY:The Agent Had a Plan—So Did I: Top Attacks on OWASP Agentic AI Sy
 stems - Parth Shukla\, Nagarjun Rallapalli
URL:https://pretalx.com/bsidesluxembourg-2026/talk/UGKRML/
END:VEVENT
END:VCALENDAR