chore: bump v2.0.0-alpha.22 — Grok Build como engine + --llm por run - #56
Merged
Conversation
Fecha o ciclo de release da PR #55: o engine `grok` (assinatura xAI, reconhecido por ~/.grok/auth.json) e a flag --llm de escopo por run precisam de uma versão nova para o npm publish. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01QiJUQ9UtghDyy6qixym8uW
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Fecha o ciclo de release da #55.
O que entra nesta versão:
grok(Grok Build, assinatura xAI): reconhecido automaticamente por~/.grok/auth.json— nunca porXAI_API_KEY, que é removida do env filho. Adaptador stream-json (tool calls ao vivo no viewer/TUI),--rulespara o papel do agente,--resumenas correções, effort clampado por modelo, trust de pasta concedido no primeiro run e hooks bilíngues (Claude Code + grok) nos dois gates de PreToolUse.--llm "[agent[,agent]=]<model> [<level>]"em todofda_*.mjs: troca a LLM/effort de UM run sem tocar no roster, impressa, traçada e salva com o run para o--resumemanter a mesma.npm version --no-git-tag-version(package.json + package-lock.json), sem outras mudanças.🤖 Generated with Claude Code
https://claude.ai/code/session_01QiJUQ9UtghDyy6qixym8uW