The vulnerability stems from insecure regex handling in the script sanitization process. The commit diff shows critical changes:- Original code used pp.pattern.matcher() without safeguards- Patched version wraps input in SecureInterruptibleCharSequence with match counting- Added test case specifically targets regex attack scenarios The injectInterruptionCalls() function was vulnerable because it processed untrusted input with potentially complex regex patterns (POISON_PILLS) without limiting regex engine operations, allowing attackers to craft inputs that would cause excessive backtracking. The introduction of SecureInterruptibleCharSequence and match counting in the patch directly addresses this vulnerability by enforcing operation limits.