AI Agents545 Bypassing EDR with Local AI
Eddie Zhang / Project Black: Claude refuses an LSASS dumper; DeepSeek builds one EDR still catches; uncensored Qwen 3.8 27B on two…
oxfemale·
Oct 2·
12 min
·545
// AI Security Research
12 postsAI Agents545 
AI Agents468 Disposable Tooling: Building LLM-Generated Mythic Agents from Prompt to Deployment
Adam Chester’s SpecterOps research on disposable Mythic agents: from a vibe-coded abomination to a two-hour Oracle harness that one-shots stage-0 implants in…
oxfemale·
Sep 21·
38 min
·468
AI Security Research689 Beyond Prompt Injection: Hacking Apple’s Private Cloud Compute (CVE-2026-20685)
A path traversal in darwin-init, the PID 1 process that provisions every Apple Private Cloud Compute node, gave a researcher an arbitrary…
oxfemale·
Aug 14·
20 min
·689
AI Security Research655 AI-Assisted Fuzzing: Generating libFuzzer Harnesses with a Local LLM
Writing the harness is the friction that keeps most codebases from ever being fuzzed. This walkthrough puts an offline open-weights model behind…
oxfemale·
Aug 13·
44 min
·655
AI Security Research800 Reverse Engineering CVE-2026-2796: How Claude Built a Firefox WebAssembly Exploit
Anthropic's Frontier Red Team showed Claude Opus 4.6 building a working browser exploit for CVE-2026-2796, a type-confusion bug in Firefox's SpiderMonkey WebAssembly…
oxfemale·
Jul 17·
15 min
·800
AI Security Research550 Disposable Tooling: Generating Mythic C2 Agents End-to-End With LLMs
A walkthrough of how an LLM, wrapped in a tiered build-and-test harness, can take a Mythic C2 agent from a one-paragraph spec…
oxfemale·
Jul 1·
12 min
·550
AI Security Research479 How LLMs Actually Work: A Transformer Internals Walkthrough
0xkato's walkthrough of the mechanisms inside modern transformer-based LLMs — tokenization, embeddings, Rotary Position Embeddings, attention with Q/K/V and causal masking, multi-head…
oxfemale·
Jun 22·
29 min
·479
AI Security Research522 Autonomous Vulnerability Hunting with MCP: Inside a Self-Improving 0-Day Pipeline
Andy Gill's ZephrSec write-up of an autonomous vulnerability hunting system built around Claude Code and the Model Context Protocol: 8 MCP servers,…
oxfemale·
Jun 19·
26 min
·522
AI Security Research338 From Prompt to Pwned: Chaining LLM and Web Bugs into Admin Takeover
A Quarkslab red team exercise stitches insecure LLM output handling, missing cookie flags and a conversation-sharing IDOR into a single-click admin account…
oxfemale·
Jun 17·
9 min
·338
AI Security Research325 CuSafe — Catching Memory Corruption on Commodity NVIDIA GPUs (USENIX Security ’26)
CuSafe is a GPU memory sanitizer for off-the-shelf NVIDIA hardware. It combines pointer tagging with in-band buffer bounds via an LLVM 21…
oxfemale·
Jun 15·
25 min
·325
AI Security Research329 AI vs Reverse Engineering: How Claude Helped Find a radare2 0-Day
Researchers used Claude AI to discover a command-injection 0-day in radare2. A crafted PDB symbol name allowed arbitrary command execution when a…
oxfemale·
Apr 10·
11 min
·329
AI Security Research391 MAD Bugs: Claude Wrote a Full FreeBSD Remote Kernel RCE with Root Shell (CVE-2026-4747)
Researchers showed that the Claude AI model could generate a working exploit for a FreeBSD kernel vulnerability (CVE-2026-4747), producing a remote root…
oxfemale·
Apr 1·
30 min
·391