nx_brand_manage.nx

nx_brand_manage.nx -- R4 MULTI-ACTOR brand management, no lock-in. Operator: "users + developers + AI-users (of Claude) + local-LLM able to manage in lots of ways to NOT lock in." The unifying mechanism: ONE brand-OPS language (line-based DATA any actor emits) applied through ONE governed mutator over the token SSOT: set|<group>|<name>|<value> -- set/add a token pack|<id> -- load a whole brand pack as the working base TWO modes: lenient (dev/CLI: skip bad ops) and STRICT-GUARDED (the untrusted-proposer seam AI/local-LLM use: any unsafe/unknown op rejects the WHOLE proposal, and the RESULT must still pass the R2 brand guard -- so a model can neither inject (<script>/breakout) nor break the brand (sub-AA)). Because the mutator is PURE DATA->DATA, the SAME ops yield BYTE-IDENTICAL output regardless of which actor sent them = no vendor/tool lock-in, and the base is untouched = reversible by construction. REUSE: nx_brand_tokens (bt_set/bt_lookup), nx_brand_guard (bg_pair_contrast), nx_brand_packs (bp_pack/bp_decode), nx_nishi_builder (nb_streq). 100% sovereign. license_tier: ORIGINAL