What problem does it solve?
During authorized penetration testing, a write-what-where primitive (format string %n, heap metadata corruption, OOB write, UAF) does not directly give control of execution. This Skill provides the methodology for selecting the right overwrite target and turning that write primitive into hijacked control flow.
Core Features & Use Cases
- Target Selection by Mitigation: Chooses between GOT/PLT entries, .fini_array/.dtors, __malloc_hook/__free_hook, __exit_funcs, and __printf_arginfo_table based on checksec results (Partial vs Full RELRO) and glibc version.
- Mitigation-Aware Guidance: Explains how Full RELRO kills the GOT path, how glibc >= 2.34 removed the malloc/free hooks, and how PTR_MANGLE protects atexit handlers requiring a pointer_guard leak.
- Confirmation Workflow: Uses a sentinel overwrite plus a gdb breakpoint to prove the hijacked call actually reaches the attacker's address before finalizing the exploit.
- Use Case: A CTF player or pentester has a format-string %n write on a Partial RELRO binary; the Skill walks them through overwriting strlen@got with system, triggering a call with "/bin/sh", and verifying the shell.
Quick Start
Ask the agent to convert my arbitrary write primitive in the target binary into code execution, selecting the overwrite target based on checksec output and glibc version.