BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//pretalx.com//bsidesluxembourg-2026//talk//8ACVB3
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-2026-8ACVB3@pretalx.com
DTSTART;TZID=CET:20260508T133000
DTEND;TZID=CET:20260508T173000
DESCRIPTION:Event Strategy & Structure\n\nCore Mission: A 2-day\, open-floo
 r "village" dedicated to exploring real-world security risks in Agentic AI
 \, Model Context Protocol (MCP) architectures\, and LLM workflows.\n\nAlig
 nment: All content and threat models are strictly aligned with OWASP guida
 nce (LLM Top 10 & AI Security Exchange).\n\nDynamic Flow: Unlike tradition
 al linear training\, this is an exploratory space. The schedule is fluid\;
  organizers will pivot topics\, attack scenarios\, and deep dives in real-
 time based on what attendees find most interesting.\n\nVillage Logistics\n
 \nOpen Access: The village runs continuously for two days with no fixed st
 art/stop times.\nDrop-in Format: Attendees are free to enter\, observe\, l
 eave\, and return at will. This supports the casual\, "hallway con" cultur
 e of BSides events.\nParallel Tracks: Multiple activities (demos\, labs\, 
 discussions) happen simultaneously\, allowing for natural scaling of depth
  from beginner to advanced levels.\n\nOrganizer Responsibilities (The Blue
  Team/Red Team)\n\nLive Operations: Organizers act as facilitators\, maint
 aining intentionally vulnerable infrastructure (LLMs\, RAG pipelines\, Aut
 onomous Agents\, MCP Servers).\n\nInteractive Walkthroughs: Instead of for
 mal talks\, organizers provide short\, continuous breakdowns of attacks\, 
 explaining why a specific trust boundary failed or how a design choice cre
 ated a vulnerability.\n\nAdaptive Defense: Based on audience feedback\, or
 ganizers will live-patch systems or remove mitigations to demonstrate how 
 security controls impact attack feasibility.\n\nAttendee Experience (The R
 ed Team)\nHands-on Exploitation: Attendees can directly interact with depl
 oyed systems to attempt prompt injection\, logic-based attacks\, and tool 
 abuse.\nFeedback Loop: Attendees actively shape the curriculum by voting o
 n which systems to attack next or requesting deeper focus on specific fail
 ure modes.\nCollaborative Defense: A key component is discussing defenses\
 ; attendees can propose architecture changes or guardrails\, which organiz
 ers can discuss or implement live.\n\nHands-on Labs & Infrastructure\nSelf
 -Paced Playgrounds: Dedicated stations will run continuously for independe
 nt learning.\nDreadnode Crucible: Focuses on practical exploitation of LLM
 s and agents.\nLakera Gandalf / Agent Breaker: Gamified challenges coverin
 g prompt injection\, goal hijacking\, and instruction drift.\nPurpose: The
 se labs ensure that even if the live demo is advanced\, beginners have a p
 lace to start learning fundamentals.\n\nAgenda: \n\nBreaking LLM Systems\n
 Theme: Fundamentals of LLM vulnerabilities and the OWASP LLM Top 10.\nLive
  Targets: Minimalist LLM deployments and chat interfaces.\nDeep Dives:\nGu
 ardrails: Examining internal mechanics and demonstrating how to bypass pra
 ctical limitations.\nRAG Security: attacking Vector Databases and poisonin
 g retrieval contexts (RAG-specific threats).\n\nAgenda: Agentic AI & MCP S
 ecurity\nTheme: The core focus of the village—Autonomous Agents and the 
 Model Context Protocol (MCP).\nComplex Workflows: Demos will feature multi
 -step agents that can plan\, execute\, and interact with external tools.\n
 Key Attack Vectors:\nInstruction Hijacking: Forcing an agent to deviate fr
 om its original goal.\nTool Abuse: Exploiting over-privileged MCP capabili
 ties (e.g.\, an agent with unrestricted file access).\nTrust Boundaries: A
 nalyzing failures in the handshake between Agents and MCP servers.
DTSTAMP:20260412T024640Z
LOCATION:IFEN room 2\, Workshops and AI Security Village  (Building D)
SUMMARY:AI Security Village - Parth Shukla\, Nagarjun Rallapalli
URL:https://pretalx.com/bsidesluxembourg-2026/talk/8ACVB3/
END:VEVENT
END:VCALENDAR
