BEGIN:VCALENDAR
VERSION:2.0
PRODID:Data::ICal 0.24
BEGIN:VEVENT
DESCRIPTION:   'Title: Sold Out - Building Agentic Reverse Engineering "Ski
 lls"\n   Tags: DEF CON Workshop | DEF CON Workshops\n   When: Saturday\, A
 ug 8\, 14:00 - 17:59 PDT\n   Where: LVCCW Level 2 W231 (Workshops) - [1]Ma
 p\n\n   Description:\n\n   Agentic reverse engineering blends interactive 
 binary analysis with\n   autonomous agent workflows. Building on workshops
  at REcon and DEF CON\n   Singapore\, this session introduces Agent Skills
 \, structured bundles of\n   instructions\, scripts\, and resources that c
 oding agents discover and\n   execute. Skills enable multi-step RE tasks w
 ith high accuracy and\n   minimal prompting via workflow capture and progr
 essive disclosure.\n\n   Participants learn how coding agents operate thro
 ugh iterative loops\n   (generate\, execute\, inspect\, refine) and how Sk
 ills plug into these\n   loops. The workshop is hands-on: attendees build 
 a multi-platform\n   driver-analysis Skill automating IOCTL enumeration\, 
 dispatch-flow\n   analysis (Windows IRPs\, Linux file ops\, macOS IOKit)\,
  code-flow\n   analysis\, and workflow capture. A capstone challenge has p
 articipants\n   build a second Skill from scratch.\n\n   Supports Claude C
 ode\, OpenCode\, Mistral Vibe\, and pi. The instructor\n   provides LLM in
 ference for all attendees\, so no paid API keys are\n   required. Students
  may also use free inference tiers from OpenCode or\n   similar providers.
 \n\n   Attendees leave with practical experience to implement agentic RE\n
    Skills in their own workflows. Basic familiarity with RE concepts and\n
    a laptop with a coding agent installed is all that is needed.\n\n   Spe
 akerBio:  John "clearbluejar" McIntosh\n\n   John McIntosh (@clearbluejar)
  is a security researcher and founder of\n   ClearSecLabs\, specializing i
 n reverse engineering\, vulnerability\n   research\, and AI-assisted binar
 y analysis. He is the author of\n   ghidriff\, an open-source Ghidra-based
  binary diffing engine\, and\n   pyghidra-mcp\, a headless Ghidra MCP serv
 er enabling LLM-driven\,\n   project-wide\, multi-binary reverse engineeri
 ng workflows. An active\n   contributor to the Agent Skills ecosystem\, Jo
 hn bridges deterministic\n   analysis with AI-driven reasoning to accelera
 te vulnerability\n   research. He has delivered training and workshops at 
 DEF CON\, Black\n   Hat\, REcon\, Ringzer0\, 44CON\, Objective by the Sea\
 , and Insomni'hack\,\n   covering topics from practical Windows reverse en
 gineering to building\n   private local LLM RE stacks. His recent work inc
 ludes the "Agentic RE"\n   training at DEF CON Singapore 2026\, the MCP Gh
 idra workshop at REcon\n   2025\, and the "Supercharging Ghidra" LLM works
 hop at Ringzer0\n   COUNTERMEASURE 2025. With over a decade of offensive s
 ecurity\n   experience\, John publishes detailed research on reversing CVE
 s\,\n   building RE tooling\, and agentic patch diffing at\n   clearblueja
 r.github.io. His teaching emphasizes reproducibility\,\n   progressive ski
 ll-building\, and contributor empowerment.\n\n   Links:\n       Registrati
 on (July 14\, 2026\, Noon US Pacific) - [2]https://events.humanitix.com/sa
 t_pm_ws6_4009\n   '\n\n   1. #LVCCW_Level2_West\n   2. https://events.huma
 nitix.com/sat_pm_ws6_4009\n\n\n
DTEND:20260809T005900Z
DTSTART:20260808T210000Z
LOCATION:DEF CON Workshops - LVCCW Level 2 W231 (Workshops)
SUMMARY:Sold Out - Building Agentic Reverse Engineering "Skills"
END:VEVENT
END:VCALENDAR