CAIN-42 CAIN Studio
25CAIN-42 Top 25 › Policy Red-Team › Canonicalization

Canonicalization

NameWhat the gate does first
decodePercent-decode path and URL arguments until stable (at most 4 rounds)
unicodeNFKC; tag characters read as the ASCII they hide; every invisible format character removed (since 2026-11-02, all of them, not six)
pathBackslashes to /, cut at NUL, resolve ../, and since 2026-11-02 /proc/self/root and friends
urlMatch on parsed scheme, host, port and path; since 2026-11-02 an IP address has one spelling
shellMatch each command segment after quote removal; refuse substitution in allow rules, and since 2026-11-02 treat it as denied in deny rules
sqlRemove comments (but keep MySQL's executable /*!..*/ since 2026-11-02), collapse whitespace, ignore case

FIRME AI CORPORATION · CAIN-42 is a Byzantine trust fabric. Numbers marked "our own check" were run by us, not by an independent party. Nothing on these pages is a certification.