code wiki / (root) / nx_adversarial_research.nx

nx_adversarial_research.nx

buildroot/runtime/nx_adversarial_research.nx

4616 B52 linesdepth 24pulls 155 transitivereach 0 importersview sourcekind tooltopic adversarial
docsdependenciesstructsconstsfunctions

about

nx_adversarial_research.nx -- THIN structured ADVERSARIAL-VERIFICATION research source organ. COMPOSES nx_research_engine (sovereign fetch->extract->bank). Operator 2026-07-02: "we need the nishi teams adversarial work to get to s-class exceed ... use the nishi researcher + census to hammer down what we need." The signed_cookie forgery bug (caught only by an ADVERSARIAL gate that tried to forge) proved the thesis: a green build that ships a forgeable MAC is worse than none -> adversarial verification is a REQUIREMENT, not a nicety. So: ground a MEASURED adversarial-exceed census on REAL sourced practice (the SOTA that fuzzing/formal-methods/red-team embody as of 2026) -- never assumptions (Rule 4). Banks clean .txt -> knowledge/library/ (indexed by nx_library_harvest_v2). license_tier: ORIGINAL

dependencies 1 imports · 0 importers

nx_research_engine.nx nx_adversarial_research.nx

imports: nx_research_engine.nx

imported by: nobody (leaf or entry point)

call flow from main pre-order; caps 40 nodes / depth 6 declared; ↻ = already shown

main rf_init nx_trust_store_load_from_c sys_mmap sys_read_file sys_openat_rd sys_lseek sys_mmap ↻ sys_read sys_close nx_nss_certdata_parse sys_mmap ↻ _pat_class_cert _pat_value_octal _pat_end _find_newline _starts_with _parse_octal_line sys_mmap ↻ _is_space _decode_octal_escape _is_octal trust_store_alloc sys_mmap ↻ nx_x509_trust_store_load sys_mmap ↻ x509_parse sys_mmap ↻ asn1_cursor_init asn1_expect_tag sys_mmap ↻ asn1_read_tlv_header asn1_read_length x509_read_tlv sys_mmap ↻ asn1_expect_tag ↻ x509_read_alg_id sys_mmap ↻ asn1_expect_tag ↻ trust_store_add

structs

none

consts

9const K_MAGIC_8388608: i64 = 8388608

functions

11func main() -> i64