ExploitBench: A Capability Ladder Benchmark for LLM Cybersecurity Agents

Exploitation is not a binary event. It is a ladder of acquiring progressive capabilities, from executing a single buggy line of code to taking full control of the target. However, existing LLM security benchmarks treat a crash as exploitation success. That single binary outcome collapses the hard...

Chrome V8 Genesis::InitializeGlobal Bugs

Chrome: V8: Bugs in Genesis::InitializeGlobal Bug: The Genesis::InitializeGlobal method initializes the constructor of RegExp as follows: // Builtin functions for RegExp.prototype. Handle regexpfun = InstallFunction global, "RegExp", JSREGEXPTYPE, JSRegExp::kSize + JSRegExp::kInObjectFieldCount...