nx_shell_bench.nx

nx_shell_bench.nx -- HONEST measured benchmark: nx_shell vs the shell competition (bash/zsh/fish/PowerShell), grounded on the fetched sh_*.raw, ADVERSARIAL/ASTERISK-disciplined (no wave). The competition is mature + huge; nx_shell is tiny + new. So the exceed is SCOPED to a sovereign minority path, and we ADMIT where we lose. EXCEED (sovereign niche, grounded): 1 ZERO-EXTERNAL-BINARY-DEPS builtins in-process (echo/grep/cd/true/false); a unix shell shells out to GNU coreutils (grep/sed/cat/ls... sh_coreutils.raw) -- nx_shell forks 0 external binaries for them. 2 ATTACK-SURFACE no env-var function-import (the Shellshock vector CVE-2014-6271, sh_shellshock.raw) and no arbitrary eval -> immune to that whole RCE class BY CONSTRUCTION. 3 SOVEREIGN-BUILD nx_cc->nxasm, no libc/gcc, 0 dynamic deps; runs Nishi organs natively (ecosystem). 4 NEVER-BRICK spawns userspace, writes 0 firmware -> brick-safe by construction (cardinal 26). PARITY: exec model (fork/execve/wait) ; pipelines+capture + control flow (; && ||). BEHIND (the asterisk -- the competition WINS here): features/scripting (job-control/history/completion/glob/ arrays/functions/expansion, sh_comparison.raw) ; maturity/ecosystem ; interactivity (line-edit/prompt/complete). *** MINORITY-PATH ASTERISK: bash/zsh/PowerShell are the MAINSTREAM and win DECISIVELY on features/scripting/ ecosystem. nx_shell EXCEEDS only on the sovereign/minimal/security niche. We are NOT a better general shell. *** T1 EXCEED=4 (each a grounded sovereign axis). T2 (teeth) BEHIND>=3 (we ADMIT losses -- no wave). T3 the MINORITY-PATH asterisk is carried. expect_exit: 0 Sovereign: nx_syscalls.