Published to npm as @bitkyc08/opencodex@2.10.2-preview.20260806 with dist-tag preview.
New Features
- Add routing profiles and why-this-route views (#1018)
- Add Command Code OAuth provider (#1066)
- Drop gpt-5.4-mini from the default intercept source models (#1089)
- Raise the management read entry cap from 200k to 500k (#1099)
Bug Fixes
- Gui: Use official brand marks for commandcode and openai icons; Shadow-call selects save the canonical namespaced model id (#1075) (#1074, #1087)
- Advertise the ladder each V4 model actually honors (#1057) (#1069)
- Strip images for the Zen models that reject them (#1043) (#1070)
- Bound the native-profile teardown and stop parsing partial JSON (#1061) (#1071)
- Warn when a startup write leaves an app-server stale (#1046) (#1072)
- Give the bounded JSON read a first-byte deadline distinct from inter-chunk inactivity (#1065) (#1088)
- Harden credential and runtime trust boundaries (rebase of #916) (#936)
Documentation
Chores
- Harden PR readiness gate — consolidated comment, draft-review trigger, G.UI waiver (#1094)
Changelog
Full Changelog: v2.10.1...v2.10.2-preview.20260806
- #936 fix: harden credential and runtime trust boundaries (rebase of #916) @lidge-jun
- #1018 feat(gui): add routing profiles and why-this-route views @Wibias
- #1066 feat: add Command Code OAuth provider @hanbinnoh
- #1069 fix(deepseek): advertise the ladder each V4 model actually honors (#1057) @lidge-jun
- #1070 fix(zen): strip images for the Zen models that reject them (#1043) @lidge-jun
- #1071 fix(test): bound the native-profile teardown and stop parsing partial JSON (#1061) @lidge-jun
- #1072 fix(codex): warn when a startup write leaves an app-server stale (#1046) @lidge-jun
- #1074 fix(gui): use official brand marks for commandcode and openai icons @Wibias
- #1087 fix(gui): shadow-call selects save the canonical namespaced model id (#1075) @lidge-jun
- #1088 fix(responses): give the bounded JSON read a first-byte deadline distinct from inter-chunk inactivity (#1065) @lidge-jun
- #1089 feat(shadow-call): drop gpt-5.4-mini from the default intercept source models @lidge-jun
- #1094 ci: harden PR readiness gate — consolidated comment, draft-review trigger, G.UI waiver @Wibias
- #1097 docs(devlog): 260806 ten-item disposition sweep + #1090 CRLF external-provider regression test @lidge-jun
- #1099 feat(usage): raise the management read entry cap from 200k to 500k @lidge-jun