Archives

A deep reference on Linux SLUB allocator internals for kernel 6.6 LTS -- covering data structures, allocation tiers, freeing paths, freelist hardening, and drain mechanics with source code references.

A research look at language-level runtime guards for AI agents — what can be caught inside Python and Node.js before a syscall ever fires, and what still has to fall through to the sandbox.

A deep dive into sandboxing AI agents on Linux — namespaces, seccomp, capabilities, and a 3-ring isolation model to keep autonomous code from going rogue.

Explores why Mixed Boolean-Arithmetic (MBA) obfuscation makes code tricky to analyze and reverse.

Announcing ir2re — my personal space for exploring, writing, and sharing