code wiki / _hdl_build / nx_game_redteam_gate.nx
nx_game_redteam_gate.nx
buildroot/runtime/_hdl_build/nx_game_redteam_gate.nx
about
nx_game_redteam_gate.nx -- ADVERSARIAL audit of MY OWN critics. A critic I calibrated to pass my own
"rich" frame and fail my own "toy" frame proves nothing; it is circular. Rigor = try to FOOL it. Two
attacks a real instrument MUST resist:
ATTACK 1 (visual): pure RANDOM NOISE. It has maximal palette + edge density + coverage. If
nx_game_critic scores it RICH/SOTA, the critic rewards BUSYNESS, not coherent art -- my "RICH 684"
game claim is then weak (measures pixel churn, not quality).
ATTACK 2 (depth): a COIN-FLIP outcome table (random win/loss, ZERO skill). It has spread + a loss +
variance. If nx_game_depth scores it DEEP, my critic conflates "has a lose state + variance" with
"skill matters" -- my "DEEP 1000" claim is then weak (a no-skill game scores DEEP).
This gate PASSES only if the critics RESIST both attacks (noise->TOY, coin-flip->SHALLOW). A RED here
is an honest finding about MY tools, reported as such -- not swept under a green rug.
license_tier: ORIGINAL expect_exit: 0
dependencies 4 imports · 0 importers
imports: nx_syscalls.nxnx_game_critic.nxnx_game_depth.nxnx_game_raster.nx
imported by: nobody (leaf or entry point)
call flow from main pre-order; caps 40 nodes / depth 6 declared; ↻ = already shown
structs
| none |
consts
| 18 | const W: i64 = 400 |
| 19 | const H: i64 = 240 |
functions
| 21 | func ww(s: *u8) -> i64 { var n: i64=0; while s[n]!=(0 as u8){n=n+1} sys_write(1,s,n); return 0 } |
| 22 | func wn(v: i64) -> i64 |
| 32 | func main() -> i64 |