Community posts

Claude Code 2.1.277 → 2.1.278 https://github.com/anthropics/claude-code/blob/main/CHANGELOG.md 277 ・CLAUDE.mdが無い場合にAGENTS.md対応 ・Plugin再インストール時の破損修正 ・Resume時のPrompt Cache/Thinking消失改善 ・大量検索やメモリ不足時の誤動作修正 ・高速入力や長時間Sessionのクラッシュを修正 ・Subagentを別AIからの出力と明示し、混同防止 ・VS CodeにTask表示Copy使用コスト表示など追加 278 ・Auto modeの安全判定をサーバー側へ移行 ・判定処理の追加料金を発生させない方式へ変更 ・/statusでサーバー判定中か確認可能 277でAgent・Plugin・Resumeを大規模安定化し、278でAuto modeの安全判定コストまで整理 ANGENTS.md対応が特に大きい #ClaudeCode #Claude #Anthropic #AI

Jev Model Routerで Claude Codeで使うモデルを 毎回自動で振り分けられるようになった‼️　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　 既に20万円以上費用を削減できてて このModを入れると Claude Codeに投げたリクエストごとに Jevが内容を見て、自動で判断します。 ・どのサブエージェントモデルを使うか ・どのメインモデルを使うか ・どのEffortレベルにするか つまり人間が毎回、 「これは軽いから安いモデル」 「これは難しいから強いモデル」 「ここはHighで」 「このタスクはサブエージェントに回して」 と考える必要がなくなる。 Jevがタスク内容を見て、 最適なモデル構成へルーティングしてくれる。 しかも、 @typesafeai API または @vercel AI Gateway 経由で利用可能。 ポイントは、 メインモデルの切り替えはセッション 開始時を中心に行うこと。 途中で頻繁に切り替えると キャッシュが壊れる可能性があるため、 そこまで考慮された設計になってる。 インストールは1コマンドだけ 👇 npx claude-code-templates@latest --mod productivity/jev-model-router これからClaude Codeは、 「どのモデルを使うかを人間が考える」 から、「やりたいことだけ伝えて、モデル選定はAIに任せる」 方向にどんどん進んでいく。 特に、 高性能モデルのクォータを節約したい人 APIコストを抑えたい人 複数モデルを使い分けてる人 Claude Codeを長時間回してる人 にはかなり相性が良さそう。 Jev Model Router、 モデルルーティングまで自動化したい人は 試す価値あります。 https://x.com/dani_avila7/status/2101176629745561686/video/1

Anthropic just adopted OpenAI's standard, a year after refusing the most-requested feature in Claude Code's repo. OpenAI already runs its agents on Anthropic's protocol. AI's two biggest rivals each won exactly one standards war, and both wars were decided by the same force. It wasn't model quality. AGENTS.md started as an OpenAI project in August 2025. A plain markdown file that tells any coding agent how to build and test your repo. Codex read it, then Copilot, Cursor, Gemini CLI, Jules, Windsurf, Zed. Over 60,000 repositories carry one today. Claude Code was the last major holdout. MCP went the opposite way. Anthropic shipped it in November 2024 as the standard for connecting agents to tools, and by March 2025 OpenAI adopted it across its Agents SDK and ChatGPT. So how did each company win one and lose one? A standard wins wherever the switching cost piles up. MCP piles up in servers. Once thousands of tool integrations existed, refusing MCP meant rebuilding all of them yourself. AGENTS.md piles up in git repos. Once 60,000 codebases carried the file, refusing it meant every one of them worked slightly worse in your product. Anthropic couldn't out-stubborn 60,000 repos. OpenAI couldn't out-stubborn thousands of MCP servers. Both companies folded to the same opponent, an installed base neither of them controls. Both standards even ended up in the same building. The Linux Foundation took in AGENTS.md in December 2025, right alongside MCP. And look at how Anthropic actually joined. AGENTS.md only gets read when there's no CLAUDE.md, shipped as a toggleable mod. That's how you adopt a rival's standard without surrendering your own file. Comply at the edge, keep the moat at the core.

Introducing Jev Model Router for Claude Code This Claude Code Mod lets you use Jev through its direct @typesafeai API or @vercel AI Gateway With every request you send to Claude Code, Jev classifies the subagent model, main model (only at session start to avoid breaking the cache), and effort level Install it with one command: npx claude-code-templates@latest --mod productivity/jev-model-router Full Mod details: http://aitmpl.com/component/mod/productivity/jev-model-router

Anthropic 终于给 Claude Code 加上了 AGENTS.md 支持。 2.1.277 版本开始，如果项目里没有 CLAUDE.md，Claude Code 会自动读取 AGENTS.md；用户也可以在 /config 里改成两份都读。AGENTS.md 是跨工具通用的项目规则文件，Codex、Cursor、Gemini CLI、GitHub Copilot 等早已支持。 这正是 Shopify CEO Tobi Lütke 八月底公开点名的问题。Shopify 同时用 Claude Code、Codex 等多种 Agent，之前 Claude Code 只认 CLAUDE.md，团队就得维护两套规则。大型代码库里只要某一层漏了同步，不同工具就可能拿到不同规则。Tobi 当时甚至说，正在考虑在 Shopify 禁用 Claude Code。 当时 Claude Code 工程师 Thariq Shihipar 还解释，Anthropic 坚持自己的 CLAUDE.md，是因为不同模型适合不同的 prompt、Skills 和项目指令格式。 不到一个月后，Anthropic 已经把 AGENTS.md 支持做成 Claude Code 的内置 Mod，并公开源码。以后开发者还可以基于这套 Mod 自己改项目指令的加载方式。

Claude Modについてメチャクチャ解説してくれてる記事発見 https://zenn.dev/nogu66/articles/claude-code-function-hooks-claude-mods

终于等来 Claude 支持 AGENTS.md 了，可以删好多代码！感觉可以通过自定义mods来支持 .agents/skills

엄청난 소식! Claude Code 이제 AGENTS.md 공식 지원하기 시작함!! 껴 자, 몇시간 전에 commit 된거,, https://github.com/anthropics/claude-code/tree/main/mods/agents-md Claude Code에서 폴더에 CLAUDE.md가 없으면 AGENTS.md를 읽고 사용함. /config에서 이 동작을 바꿀 수 있구요. 이제 이 기능이 Claude Code mods(=곧 공개될 하네스 커스터마이즈 방식) 위에 만들어진 내장 모드. 지금은 기본 제공이지만, 나중에 프로젝트 지시문을 직접 커스텀하는 모드도 만들 수 있게 됨. 아시다시피 AGENTS.md는 Codex, Cursor, Copilot, Gemini CLI 등 여러 코딩 에이전트가 읽는 벤더 중립 프로젝트 지시문 파일이라고 생각하시면 됨.

Do you like Claude Code's auto tool approval mode? @sarahwooders made a mod that uses Jev from @typesafeai to your Letta agent!

シルバーウィークのお供に、『Claude Mods』に入門してみませんか？ 画面表示の変更や独自ツールの追加など、Claude Codeハーネス深く拡張できる仕組みです すでに、入門している方はどう使っているか教えて欲しいです！ https://zenn.dev/nogu66/articles/claude-code-function-hooks-claude-mods

Claude Code 终于支持 'AGENTS.md' 了 从 2.1.277 版本起，Claude Code 原生支持 AGENTS.md。规则很简单：当目录下没有 CLAUDE.md 时，Claude 会自动查找并读取 AGENTS.md 作为替代的项目指令文件。该行为可在 /config 中开关。 从 cc 的开源仓库的实现来看，实际上共有四种模式：仅 CLAUDE.md（不启用）、回退到 AGENTS.md（默认）、两者同时加载并去重、仅使用组织托管文件。 通过 Mods README 能看到 AGENTS.md 与原生 CLAUDE.md 的差距：@ 提及、IDE 选区、notebook/图片/PDF 读取时不会附加 AGENTS.md；不支持符号链接解析；/memory 界面看不到它。https://github.com/anthropics/claude-code/tree/main/mods/agents-md # 最后看看 @trq212 提到的 “基于 Mods 机制构建” 是什么意思？ 官方说明：A mod is a Claude Code plugin whose behaviour lives in a hooks module mod 就是一个 Claude Code 插件，其全部行为位于一个 hooks 模块中：一个 register(on, options) 入口，把引擎事件挂载为 ($, e, next) 形式的函数。Claude Code 内置了四个 mod。 一个 mod 长什么样： · .claude-plugin/plugin.json：插件清单 · hooks/hooks.json：声明入口模块 · hooks/ 下的 TypeScript：具体实现，类型来自 types/ 目录的声明 · tests/：测试，用 claude plugin test mods/<name> 运行 目前内置的四个 mod 及其职责： 1. sec-default：在托管设备或 Team/Enterprise 组织环境下，保护组织的 hooks、managed settings 和工具策略不被用户自行安装的插件篡改，"自身不添加任何策略"，位于最外层。 2. diff：提供 /diff 命令，在对话旁的面板中逐文件、逐 hunk 地展示未提交的会话改动，随 Claude 编辑文件实时刷新。 3. telemetry：在 engine.create 阶段向 $ 对象注入 $.telemetry（log、mark），让其他插件能记录分析事件；分析关闭时不发送任何内容。 4. agents-md：让 AGENTS.md 作为项目指令文件被加载，即本次的功能。

Claude Codeが「AGENTS.md」に ネイティブ対応した！！ これ、地味に見えて AI開発の歴史的な転換点です。 AIコーディングツールを 併用してる人なら分かるはず。 ︎︎︎︎︎︎☑︎CursorやCodexは「AGENTS.md」 ︎︎︎︎︎︎☑︎Claude Codeは「CLAUDE.md」 これまでは同じ開発ルールを 別ファイルで二重管理するか CLAUDE.mdから `@AGENTS.md` で 読み込ませる等の工夫が必要でした。 今回のアップデートで、 CLAUDE.mdがない場合は Claudeが自動でAGENTS.mdを パースして動作するように収束。 さらに見逃せないのが、 裏で進む 「Claude Code Mods」の存在です。 単なる指示ファイル対応にとどまらず TypeScriptフックでAIの動作基盤 そのものをカスタムする仕組みが 整いつつあります。 今起きているのは「AIツールの標準化」 バラバラだったAIコーディング環境が、 共通フォーマットへ 一気に収束し始めています。 この「標準化の流れ」を掴んだ人が、 今後のツール選定と開発効率で 圧倒的な差をつける。

Put together a claude code plugin to use jev with openrouter, check it out! https://github.com/sebastianbugal/jev.git

🚨 خبر عااااااجل لمستخدمي Claude Code: Claude Code أخد خطوة جديدة مهمة جدًا. دلوقتي AGENTS.md بقى مدعوم رسميًا، وده بيسهّل عليك توحيد تعليمات المشروع وتوجيه Claude لطريقة الشغل اللي أنت عايزها. والأهم إن التحديث مرتبط بنظام Mods الجديد، واللي بيفتح الباب لتخصيص Claude Code بشكل أكبر. لو بتستخدم Claude Code يوميًا في البرمجة، التحديث ده يستحق التجربة فعلًا.

Harness Learning #4: Claude mods Claude now supports AGENTS.md through claude mods. - a mod is a plugin that uses function hooks to extend claude code functionality - a mod can register many hooks: session startup, tool call, turn completion, etc. - basically register callbacks based on different agent events - hooks can chain and call one another - mods can also provide services or state to other mods. Ex: a telemetry mod registers a telemetry logging service that any other mod can use Simple example hook export function register(on) { on("turn.complete", async ($, event, next) => { const result = await next(event) $.ui.toast("Claude finished!") return result }) } Reminds me of @pidotdev @badlogicgames extensions :)

AGENTS.md support is built off of Claude Code mods, our upcoming way to customize the Claude Code harness. This is a built-in mod, but you’ll be able to build custom versions of project instructions yourself as you’d like too.

If you want to try Jev and have no idea where to start, this is basically what I did. 1. Get access from TypeSafe: http://typesafe.ai 2. Install the skill so Claude Code knows how to use it: npx skills add typesafe-ai/skills --skill typesafe-ai If you’re using the Claude plugin instead: claude plugin marketplace add typesafe-ai/skills claude plugin install typesafe@typesafe-ai 3. Create an API key from the TypeSafe dashboard and you’re pretty much ready. The fun part is figuring out where to actually use it. Don’t go into your code and replace every GPT call with Jev. That’s not really the point. Look for all those tiny decisions your agent is making. “Should I search again?” “Which tool should I call?” “Is this result useful?” “Does this need a human?” “Should I keep this piece of content?” “Is this lead worth following up on?” These are the calls I’d experiment with first. For example, if you’re building a research agent, you could have Jev look at 10 search results and simply decide which ones are worth sending to your main LLM. Or for a content agent: Jev: Is this idea worth writing about? → yes Jev: Which format? → thread Jev: Which audience? → designers Then your actual LLM does the writing. That’s the bit that clicked for me. You don’t really want Jev to be your “smartest model”. You want it sitting in all the boring little decision points around your smarter models. I’ve started experimenting with this on content workflows. Will share what I find once I have something interesting. Because honestly, that’s probably a more useful test than another benchmark screenshot.

Anthropic open sourced 11 plugins that turn Claude into a specialist for any office role They're built for Claude Cowork and work in Claude Code too. Each plugin bundles the skills, tool connections and slash commands for one job function. A few of them: > Sales: research a prospect, prep for the call, review your pipeline, draft the outreach. Connects to HubSpot, Clay, ZoomInfo. > Finance: journal entries, account reconciliation, financial statements, variance analysis, month-end close. > Legal: contract review, NDA triage, risk assessment, compliance questions. > Data: write the SQL, run the analysis, build the dashboard, validate the numbers before you share them. > Enterprise-search: one query across email, chat, docs and wikis. Also in the repo: productivity, customer support, product management, marketing, bio research, and one for building your own plugins. How to use it in Claude Code: claude plugin marketplace add anthropics/knowledge-work-plugins claude plugin install sales@knowledge-work-plugins In Cowork, just install them from the repo. They're plain markdown and JSON, so swap the connectors for your own stack and drop your company's terms and processes into the skill files. That's where they stop being generic.

Most people will use Jev wrong They'll try to replace their main model with it That's not where the value is Jev makes more sense for the tiny decisions inside your agent that happen hundreds of times Those tasks don't need Claude, GPT or another frontier model every single time Setup is simple: Get access at : https://typesafe.ai Install the TypeSafe skill npx skills add typesafe-ai/skills --skill typesafe-ai For Claude Code: claude plugin marketplace add typesafe-ai/skills claude plugin install typesafe@typesafe-ai create your API key Ask your agent to use the TypeSafe skill From there, start replacing basic decision calls with Jev The biggest improvement comes when you use it across the whole agent flow: router → pick gate → validate judge → check result You can also batch many decisions into one request and send uncertain cases back to a larger model. That's the real use case: not replacing intelligence, but removing expensive intelligence from places where you never needed it

Claude Code 2.1.275, 2.1.276 (抜粋) - Claude appsゲートウェイのサインインに、サインイン中のアカウント表示を追加。ゲートウェイがアカウントを名指しする場合、認証情報を保存する前に確認を求められるようになり、`/status`にも表示される - 現在のターンを中断してキューに入っているメッセージを一括送信する送信キー（ctrl+enter、またはctrl+x ctrl+s）を追加。送信済み・キュー中のメッセージは、モデルが受信するまで灰色で表示される - 設定済みの`otelHeadersHelper`が失敗した際の起動時警告を追加。テレメトリを何も出力していないことに気付かないセッションを防ぐ - http://claude.aiアカウントで有効化されているskillとpluginを、そのアカウントでサインインしたターミナルセッションに同期する機能を追加。`syncClaudeAiSkills: false`または`syncClaudeAiPlugins: false`でオプトアウトできる - `/plugin install <plugin> --marketplace <source>`を追加。plugin installの前にmarketplaceの追加を提案する - Artifactツールのpublishとreadの結果を改善。誰がページを開けるか、ownerのShare menuが何を提供するかを示すように - ペースト・添付した画像を改善。Desktop・VS Codeを含め、Claudeがパーミッションプロンプトなしでファイルとして開ける場所に保存されるように - Claude in Chromeのauto modeを変更し、bypass modeと同様に、classifierが承認した呼び出しについては拡張機能のper-siteチェックをスキップするようになった。リダイレクト後の`browser_batch`の「Permission denied」を修正 - [VSCode] Memoryダイアログ内で保存済みmemoryの表示・編集・削除を追加 - [VSCode] テキストを入力せずに添付画像を送信する機能を追加 - [VSCode] 提案された変更のdiffタブの各変更にaccept・rejectボタンを追加し、変更ごとにレビューできるように - [VSCode] agent mapを改善。pillは稼働中のagent数を数え、失敗後は赤くなり、メインagentはmapのスクロール中も表示され続け、agentはstateと終了時刻でソートされるように - [VSCode] Claudeが作業中に送信したメッセージが、Claudeが着手するまで会話の下部で待機するように変更 - [Claude Tag] Claude Tag設定のaccess bundleにattach conditionsを追加。Ownerは、member-onlyだけでなくguestを含むチャンネルやSlack Connectチャンネルにもbundleを適用できるように - [Claude Tag] Claude Tag admin settingsのaccess bundleのCredentialsタブに、Amazon CloudWatch・CloudWatch Logs・Amazon SNS・Google Cloud Monitoring・Cloud Loggingのpresetを追加 - [Claude Tag] US3・AP1・AP2・US1-FEDサイト向けのDatadog presetを追加。新しいDatadog接続はDatadogのreadとquery APIルートに限定されるようになった https://github.com/anthropics/claude-code/blob/main/CHANGELOG.md#21275

GROK BUILD Sources: official http://docs.x.ai/build + http://x.ai/cli + your Grok Build bookmark folder (322 posts) Updated: 2026-09-18 NEW TODAY - Folder now 322 posts (3 new vs 319 on Sep 17) - Testers (mark_k / SpaceXAI): v1.0.35 — better MCP status reporting and tool rendering; Swift syntax highlighting; cleaner copy/paste in minimal mode; improved /memory UI and much faster memory - Official changelog still Latest v1.0.34 · Sep 16, 2026 (checked Sep 18; v1.0.35 not on http://x.ai/build/changelog yet) - Bookmarks also: XFreeze on how fast Grok Build evolved; testerlabor claw-machine demo with Grok Bots WHAT IT IS - A terminal-native coding agent with a fullscreen TUI - Also runs headlessly in scripts, or inside other apps via ACP - Understands the repo, edits files, runs shell commands, and manages long jobs - Open source harness: http://github.com/xai-org/grok-build - Different from Grok Bot (cloud teammates) and Digital Optimus (real-time video control) - Bookmark testers: native /remote-control support is coming (SpaceXAI confirmed); Warp now ships built-in Grok Build CLI with /remote-control (official grok Sep 11) INSTALL AND UPDATE [] macOS / Linux / WSL: curl -fsSL http://x.ai/cli/install.sh | bash [] Windows PowerShell: irm http://x.ai/cli/install.ps1 | iex [] Then: grok --version [] Update: grok update (stable) or grok update --alpha - Testers: v1.0.5 adds GROK_CONFIG env, auto worktree reclaim, hook-block wording, RTL, ACP effort, image/video call limits - Testers: v1.0.6 changes subagent spawning (breaking) and adds a customizable status line - Testers: v1.0.7 adds a first-class workflows catalog, Always/Never rules, and a timer-refresh status line - Testers: v1.0.8 improves MCP servers and workflows. Concurrent subagents start faster without freezing the parent or UI. Follow-ups send while a task is running. Ctrl+S stashes the prompt draft. /workflow autocompletes. MCP can request form input or URL consent. Workflow rows show current context usage - Testers: v1.0.9 adds agent budget + reasoning controls, plugin-provided agents, /feedback images, instant TUI mode switching, smarter clone/worktrees, resilient subagents, faster MCP startup, concurrent agents without rate-limit collapse - Testers: v1.0.10 — grok clone reuses matching local checkouts as linked worktrees for faster session creation - Testers: v1.0.11 — Auto/headless QoL: resume-picker headless sessions, configurable default permission mode, Auto subagent messages, mkdir/touch without prompts, no 10-hour monitor timeout, faster background waits, turn duration after /resume - Testers: v1.0.12 — faster worktrees; MCP retry; subagent wait after interjection; hook descriptions; context/token accuracy; auto-compaction bar; .NET watcher fix; Auto recap during background streams - Testers/Official: v1.0.13 · Aug 28 — compressed CLI downloads; truncated responses continue; hooks can confirm (not only allow/deny); inference retry on stalls/drops/5xx; Windows ~/.grok and worktrees; more reliable session save; faster MCP startup when auth configured; truncated tool calls with complete args execute; large-image session fix; scheduled-task UUID IDs + stop reminder - Testers: v1.0.14 — PostToolUse hooks feed model feedback; grok usage <session-id> per-turn tokens/cost; proactive OIDC refresh; stronger subagents; ~70% smaller Windows CLI downloads; large-session memory no longer blocks turns - Testers: v1.0.15 — background model connect/token refresh/MCP at startup (faster first reply + new sessions); memory consolidation deferred off exit path; /copy and /export tip after repeated drag-copies - Testers: v1.0.16 — MCP servers at session bind for workspace integrations; enterprise model restrict via signed requirements.toml; 1h default wait ceiling for long subagent/task output; MCP OAuth from /mcps deadlock fix; sessions recover after expired-token network interrupts - Testers: v1.0.17 — MCP multi-round-trip elicitation (input_required → resume); quieter next-prompt ghost suggestions; /btw panel table cell select/copy as TSV - Testers: v1.0.18 — managed policy blocks disallowed MCP servers and marketplace installs before write; interactive Chart.js charts; richer image-gen params via composer state; configurable per-model rate-limit retry; per-model mTLS for secure upstream; faster startup/first responses with more work in background; session/subagent/auth/composer/UI fixes - Testers: v1.0.19 — scheduled /loop tasks always run in background (no turn inject); session resume reports running loops/subagents/workflows; headless `grok -p --worktree`; dashboard shows new sessions immediately + /usage from dashboard; transparent terminal theme; faster first prompt after login with MCP; UI freeze/crash fixes - Testers: v1.0.22 — first-party desktop MCP; `grok-workspaced daemon` for Computer Hub folder expose; continue finished subagents; background agents return real output; real line-number diffs + auto-expand on edit permission; built-in tools beat user MCP name collisions; remote-control device grouping; Auto blocks destructive `git checkout --`; faster open/resume (MCP config off critical path) - Testers/Official: v1.0.23 — bare URLs and emails inside table cells stay clickable on every wrapped line - Testers/Official: v1.0.24 — Esc no longer cancels a running turn (use Ctrl+C) - Testers/Official: v1.0.25 · Sep 9 — `/workflow` pause/stop for agent-launched runs; durable background-task snapshot across reconnect; silent successful hooks; full Bash pager output; `/theme` alias match; voice inserts at cursor; scheduled-task cleanup reminders; `/feedback` immediate send; dashboard/actions/UI fixes - Testers: v1.0.26 · Sep 10 — subagent interjects interrupt long background waits; scrollback labels how messages were sent; feedback form available in minimal mode - Testers: v1.0.27 · Sep 10 — queued messages no longer duplicate when sent immediately; dashboard spinners stay live during background tasks - Testers: v1.0.28 · Sep 10 — mid-message `/btw`; per-model `reasoning_summary` in config.toml; safer multi-process auto-update; soft-wrapped prompt Up/Down fix - Testers: v1.0.29 · Sep 11 — pager Tab cycles prompt/scrollback/whole dock; Ctrl+G hide/show dock; concurrent `grok agent stdio` same GROK_HOME startup crash fixed; MCP 2026-07-28 servers; Windows Client-ProjFS-off startup fix; idle dashboard CPU paint reduced - Testers: v1.0.30 · Sep 11 — session header single-row (overlay context + Dashboard); tmux pane lag fixed; elapsed time into hours; Watchers/Loops next-run due; workflow status clickable in dock - Testers: v1.0.31 · Sep 13 — multi-agent readability (running vs finished subagent group counts); dashboard search/resume predictability; subagent status labels; dock sections/paths; cancellation banners/messages; MCP long server-prefix tools fixed - Testers: v1.0.32 · Sep 14 — plugin/skill listings reflect latest config immediately (even before a session); Windows ARM64 first-handshake TLS crash fixed; slash commands like /feedback typed during plan approval execute correctly - Official/Testers: v1.0.33 · Sep 15 — MCP tool results include structured JSON when the server provides it; cancelled tool calls tell the server to stop; manage memory from /memory (delete with two-press x); GROK_GROVE=1 enables clone + worktree when knobs are set; long-running session + agent reliability; SSH Alt+Enter footer; macOS paste image fix; and more (see changelog) - Official/Testers: v1.0.34 · Sep 16 — Memory generally available (remembers conventions, decisions, project facts across sessions; /memory browse; /dream organize notes into topics); Markdown headings receive theme colors correctly - Testers: v1.0.35 · Sep 17 — better MCP status reporting and tool rendering; Swift syntax highlighting; cleaner copy/paste in minimal mode; improved /memory UI and much faster memory - Official changelog Latest v1.0.34 · Sep 16, 2026 (checked Sep 18; Memory GA + v1.0.31–v1.0.33; testers already on v1.0.35 not listed yet) - Bookmark testers: Grok Build mode is now on all SuperGrok and X Premium plans, web and mobile - Testers: official Netlify plugin for build and deploy - Official/Warp: Grok Build CLI runs natively inside Warp (rich prompts, /remote-control, file explorer + code review) - Testers: Quo communications plugin — SMS (individual/group/bulk), contacts, tasks, message/call transcripts, missed-call follow-up (`grok plugin install quo --trust`) - Bookmark testers: Grok Build can find/sequence footage and do real image/video editing (cut, assemble, Photoshop-style changes) from a brief — not only generate; save preferred designs/workflows as skills - Bookmark testers: Grok Build Mode can ship simple web apps (example: image watermarker) - Bookmark testers: Grok Build can drive Blender 3D workflows (examples: rocket launch site from scratch; Blender + Imagine car-chase cinematic clip) - Bookmark testers: a usage-reset control can set usage back to 0% - Changelog: http://x.ai/build/changelog - Evaluate Grok 4.6 with the Grok Build harness (http://grok.com/build) - Docs: http://docs.x.ai/build/overview - Download page: http://x.ai/cli SIGN IN [] First launch opens a browser for authentication [] Headless or remote: grok login --device-auth [] Or: export XAI_API_KEY="xai-..." then grok [] Sign out: grok logout [] In TUI: /login and /logout START A SESSION [] cd your-project then grok [] First prompts: Explain this repo. or src/main.rs Walk me through this file. [] grok -p "Explain this codebase" for a one-shot headless run [] grok -c continues the most recent session in this directory [] grok -r resumes a session by id (or the most recent) [] grok inspect shows discovered config, skills, plugins, hooks, and MCP [] grok models lists available models [] Switch model in TUI with /model LEARN THE TOOL [] Run /tutorial or /help [] /release-notes or /changelog for the current version [] Read AGENTS.md in the repo (also CLAUDE.md if you are migrating) [] Import Claude settings with /import-claude [] Carry over rules, skills, and MCP from Claude / Cursor / Codex [] Keyboard shortcuts: /help and Ctrl+. [] Voice: Ctrl+Space or /voice [] Shift+Tab cycles Plan, Auto, Always-approve HOW TO WRITE A TASK [] Outcome: what should be finished [] Sources: files, folders, URLs, or mentions [] Constraints: what not to touch, and when to ask [] Deliverable: diff, report, test, or published app [] Review point: plan first, or stop before commit/push [] For big work: /plan first, then approve [] For long ownership: /goal [] For research first: /deep-research [] For a sick laptop: tell it what is broken and let it investigate [] Testers (tetsuo): start with work you can check — name the behavior, give a reproduction command, limit files, list final checks, say where to stop [] Testers (tetsuo): inspect first (git status --short, grok inspect); give executable feedback (compiler/linter/tests) so the agent can loop; review the full diff and the command behind every green check CORE COMMANDS [] /new starts a new session [] /home returns to the welcome screen [] /resume and /sessions switch sessions [] /fork branches the current session [] /rename or /title names the session [] /session-info shows session details (click or drag to copy) [] /context shows token and skill/MCP cost [] /usage shows credits [] /memory browses session/project memories [] /dream organizes recent notes into topics [] /compact condenses history [] /rewind truncates conversation history (confirms; does not wipe files) [] /export saves the conversation [] /btw asks a side question without derailing the turn (v1.0.28: also works mid-message — text after `/btw` is the side question and stays out of the main transcript) [] /check-work has a subagent review the code [] /feedback files a report [] /doctor and grok doctor fix terminal, tmux, clipboard, keyboard [] grok du shows disk use under ~/.grok [] grok wrap runs a command with local clipboard support [] grok trace exports debug traces [] /loop runs a prompt on an interval (auto-expires after 7 days) [] Queued follow-ups can send as interjections if you turn that setting on [] Ctrl+4 opens the Agent Dashboard [] Ctrl+U can continue a Claude Code session inside Grok Build PLAN MODE AND PERMISSIONS [] /plan or Shift+Tab enters plan mode [] Only the plan file can be edited until you approve [] Auto and always-approve do not skip plan review [] a approves, s requests changes, c comments, q quits plan [] /view-plan reopens the preview [] /auto auto-approves safer tools [] /always-approve skips most prompts (deny rules and hooks still apply) [] Edit always-allow bash as free-form globs [] Restrict web search domains with [toolset.web_search] in config.toml [] PreToolUse hooks can rewrite tool input, not only allow or deny [] StopCancelled reports when a turn is interrupted or rejected [] Pin auth to API key or OIDC in config.toml if needed PARALLEL WORK AND AGENTS [] Agent Dashboard: /dashboard or Ctrl+4 [] Run feature, review, fix, and test agents from one place [] Pause, check progress, and jump into any agent [] Worktrees: grok -w or Ctrl+W (isolated git copy under ~/.grok/worktrees) [] grok worktree list / show / rm / gc [] /fork --worktree forks into a worktree [] /goal owns a long engineering objective (status, pause, resume, clear) [] /create-workflow authors a repeatable multi-agent pipeline [] /workflow launches, pauses, resumes, or stops a run [] /workflows is the live run dashboard [] /deep-research starts a built-in research workflow [] Tools and MCP servers receive GROK_SESSION_ID [] Background tasks and TODOs survive compaction VOICE AND MEDIA [] Dictate with Ctrl+Space [] /imagine generates an image [] /imagine-video generates a video [] Ask it to boost, compress, and re-encode video audio in plain English [] One-prompt 3D: ask it to model in Blender and watch it drive the app [] Morning voice-memo automations can research and read you a briefing [] Disable image/video tools via config or env if you do not want them [] Testers: save preferred image/video edit designs and workflows as skills so repeats match your style MEMORY (v1.0.34 GA) [] Grok Build remembers conventions, decisions, and project facts across sessions [] Notes capture in the background after useful work [] /memory to browse and manage memories (delete with two-press x) [] /dream to organize recent notes into topics [] The more you build with it, the more useful it becomes (official grok) SKILLS PLUGINS AND MCP [] /marketplace installs plugins [] /plugins /skills /hooks /mcps open the same extensions modal [] Skills live in ./.grok/skills, ~/.grok/skills, and plugin skill folders [] User skills appear as slash commands; use /local:name if names collide [] grok plugin install / enable / disable [] grok mcp list / add / doctor [] Type / to reference a skill [] Testers: save preferred designs and editing workflows as skills so Grok Build reuses your style every time [] Common marketplace plugins: Exa, TinyFish, MongoDB, Vercel, Sentry, Cloudflare, Chrome DevTools, Supabase, Neon, Firecrawl, Figma, Railway, Stripe [] Official/Testers: browser-use plugin — /plugins or `grok plugin install browser-use --trust`; enable Chrome remote debugging to drive your signed-in local Chrome (or isolated cloud browser; local uvx, no API key for local Chrome) [] Managed MCP servers come from the gateway catalog [] Add a local directory as a marketplace source if you keep private plugins [] Testers (v1.0.33): MCP tools can return structured JSON when the server provides it; cancelled tool calls notify the server to stop [] Official (v1.0.34): /memory browse/manage; /dream organize notes; Memory GA across sessions [] Testers (v1.0.33): GROK_GROVE=1 enables both clone and worktree features when specific knobs are set WORKTREES AND FILES [] Worktrees start from current HEAD, including uncommitted changes [] grok -w --ref main starts from a clean ref [] Ending a session does not delete the worktree; run grok worktree gc [] Keep durable files in the repo or ~/.grok, not temp dirs [] Config: ~/.grok/config.toml (Windows: %USERPROFILE%\.grok\config.toml) BUILD MODE AND PUBLISHING [] Build Mode on http://grok.com (and mobile where available) [] One prompt can become an app on a http://grok.me domain [] Export source to GitHub [] Examples: http://driver.grok.me, http://bbox.grok.me, http://gravity.grok.me [] SuperGrok Heavy may be required for some publish features [] Bookmark testers: Grok Build is now available for free users too — describe what you want, Grok builds apps/websites/games/dashboards, then publish and share [] Testers: Tesla Owners Silicon Valley — creating from a prompt is open to free users; a live http://grok.me URL still requires a paid plan [] Testers: Android (X Freeze) — GitHub push, secrets store, invite-only lock, download builds, custom domains, share to X; end-to-end ship from a phone [] Testers: Android Library (Menu > Library) — Imagine generations, Grok Build mode apps, and files created/uploaded; mobile access like desktop HEADLESS AND AUTOMATION [] grok -p "task" for scripts [] grok -p "task" --output-format streaming-json [] grok agent stdio for ACP in other apps [] grok serve --protocol acp --port 9120 for a local ACP server [] Custom models go in ~/.grok/config.toml under [http://model.name] COMMUNITY TOOLS [] GrokTerm: multi-harness terminal host with voice [] desktop-harness: Mac Accessibility control (http://github.com/xfreeze2/desktop-harness) [] Quill: universal SuperGrok voice layer on macOS [] AFK Pilot / Grok Build Desktop: remote edit and syntax highlight [] Grok Build for VS Code (Community) extension — Testers: Routines dashboard for scheduled executions across Grok Build, Codex, and Claude Code (89K+ installs; not affiliated with SpaceXAI) [] Unity CLI / MCP for game work SAFETY AND USAGE [] Prefer plan mode before large edits [] Prefer always-approve only after you trust sandbox and deny rules [] Opt in to share traces only if you want to (off by default) [] Privacy banner can be dismissed from Settings [] Video tools may require ZDR storage [] Check /usage before long multi-agent days [] Grok 4.6 may run double-token promos; check current offers OFFICIAL LINKS [] Overview: http://docs.x.ai/build/overview [] Modes and commands: http://docs.x.ai/build/modes-and-commands [] Plan mode: http://docs.x.ai/build/features/plan-mode [] Worktrees: http://docs.x.ai/build/features/worktrees [] Skills and plugins: http://docs.x.ai/build/features/skills-plugins-marketplaces [] CLI reference: http://docs.x.ai/build/cli/reference [] Changelog: http://x.ai/build/changelog [] Install: http://x.ai/cli [] Source: http://github.com/xai-org/grok-build

How to use Jev, and where it actually gives you the 100x: setup takes 10 minutes: 1. join the waitlist, people are getting approved same day 🔗 https://typesafe.ai 2. install the official skill so your agent writes correct calls: - npx skills add typesafe-ai/skills --skill typesafe-ai on Claude Code it's two commands, the marketplace add on its own doesn't install anything: - claude plugin marketplace add typesafe-ai/skills - claude plugin install typesafe@typesafe-ai 3. create an API key in the dashboard 4. in your prompt just say: "use the TypeSafe skill" now the part nobody is posting: the 100x isn't the model, it's where you put it you don't get it by swapping your LLM for Jev you get it by deleting the calls that never needed a language model open your agent and find every call that just picks something: > which tool next > is this spam > is this chunk relevant > does this need a human > is this diff risky none of those are writing tasks they're if statements you outsourced to a frontier model here's the upgrade, in order: 1. replace each one with a typed question Choice picks from up to 255 options, Score places it on a 2-10 level scale, Noul returns a raw 0-1 2. batch them questions in one call run in parallel and barely move the latency, and output tokens are free so ask every question you might need, including the ones you'll throw away 3. threshold on confidence, not on the answer under 0.5 escalate to a big model or a human 0.85+ before anything irreversible 4. never let it invent options build the candidate list in code, from the DOM, the retriever, the tool trace then let it pick 5. put it in the loop, not next to it router picks the cheap model, gate checks the tool call before it runs, judge verifies the output after that's where the heaviest calls in your agent are hiding 6. start with compaction tonight score every tool call, drop the dead ones, keep the survivors verbatim instead of a lossy summary lowest effort win available and you'll see it on tomorrow's bill the honest part: text only right now, no images, no audio and on broad benchmarks it loses to frontier models but somebody ran 18,514 emails through it zero-shot and got 98.33% against a TF-IDF classifier trained on 14,800 labelled examples that got 98.39% no training data, $1.12 total it wins on narrow, well specified decisions which is most of what your agent is actually doing all day today gonna share use case how i integrated it to content creation and how i find winning meta ads now in a seconds...

Claude Code can now compact without a summary. claude plugin marketplace add tamaratran/fast-jev-compaction Normal /compact asks the model to retell your session. Exact file paths, errors and constraints quietly die in that retelling. This plugin never rewrites anything. It only deletes what's no longer needed. How it decides: → Jev, TypeSafe's new decision model, scores every tool call and result → each one gets a keep probability, all in one fast request → stale reads, greps and test logs get dropped or cut to 300 chars → your prompts and Claude's replies stay verbatim, in order If Jev can't cut enough, it falls back to the built-in summary. Nothing breaks. And this is the part nobody mentions: The context you keep after compaction is the real context. Not a summary of it written by the same model that's running out of room. MIT. Needs Claude Code 2.1.274+, function hooks on and a TypeSafe API key. Jev is on a waitlist. Link in the comments

Claude Code plugin that captures your corrections via regex + semantic analysis, then syncs them to CLAUDE.md and AGENTS.md with confidence scores from 0.60–0.95. 160 tests, runs on macOS/Linux/Windows, no WSL. @claudeai https://github.com/BayramAnnakov/claude-reflect

さっき試したところ、Claude Codeの安全装置でブロックされて使えなくなっていました Claude Modsは実験的な機能なので、使えるようになるプラグインもあるが、逆に使えなくなるプラグインもありそう

Jev를 이용해 Claude Code 컨텍스트 순간 압축하기 (1초만에 1M->86K) 일반적인 compaction은 예전 턴을 요약합니다. 그러면 파일 경로, 정확한 에러 메시지, “이 파일은 절대 건드리지 마” 같은 제약이 사라질 수 있습니다. fast-jev-compaction은 필요한 내용, 유저의 텍스트는 건들지 않고 tool call / tool result만 삭제합니다. 나머지는 원문 그대로 유지합니다. 만약 컨텍스트가 충분히 줄어들지 않는다면, Claude Code의 기존 압축으로 fallback 됩니다. 가장 간단한 설치법은 아래 명령어를 입력하시면 되고, Claude에게 레포를 던져주셔도 됩니다. ---- export CLAUDE_CODE_ENABLE_FUNCTION_HOOKS=1 export TYPESAFE_API_KEY="여기에_TypeSafe_키" claude plugin marketplace add tamaratran/fast-jev-compaction claude plugin install fast-jev-compaction@fast-jev-compaction ---- 레포 링크 https://github.com/tamaratran/fast-jev-compaction

『 terminal-browser』を使うと、Claude Code内で直接ブラウザを開ける Chromiumの画面をkitty Graphics Protocolでターミナル内に描画しているので、kittyやGhosttyなど対応ターミナルでは本物のブラウザが動く Claude Mods の使用を前提で考えると、「kitty Graphics Protocolに対応しているか」でターミナルを選びたくなる https://x.com/RobKnight__/status/2100622380439683541/video/1

【速報】AnthropicがClaude用の職種別プラグイン11本を「knowledge-work-plugins」としてGitHubで公開した。営業、サポート、法務、経理など、職種ごとにスキルとコマンドとコネクタを1セットにまとめたもの。 ・中身はmarkdownとJSONだけ。コードもビルドも不要 ・plugin.json、.mcp.json、commands/、skills/の4点構成 ・skills/はClaudeが場面に応じて自動で読む、commands/は/sales:call-prepのように自分で叩く ・Claude Codeなら`claude plugin marketplace add anthropics/knowledge-work-plugins`で追加できる スキルを1枚ずつ書き足してきた人は、職種でまとめる置き方を一度見ておくといい。 ↓詳細

Claude Codeからissueを管理できる TUIで追加と閲覧をできる Claude Mods を作ってみた

Claude Code 2.1.275、来てました。変更96件（本体53＋VSCode27＋Claude Tag9＋Web5＋Code Review2）で、274の108件に続いて2日連続の大型回。npm公開は日本時間9/18 05:20で、274から21時間44分。272の明細は今回も1行のままです 📌 追加5件、いちばん効くのは「今すぐ送る」キー ・ctrl+enter（または ctrl+x ctrl+s）で、いまのターンを止めてキューに溜めたメッセージをまとめて送れる。送ったぶんはモデルが受け取るまでグレー表示 ・http://claude.aiアカウントで有効にしたスキルとプラグインが、同じアカウントでサインインした端末のセッションにも同期される（`syncClaudeAiSkills: false` / `syncClaudeAiPlugins: false` で切れる） ・`/plugin install <plugin> --marketplace <source>` で、マーケットプレイス追加とインストールが一手で済む ・`otelHeadersHelper` が失敗したら起動時に警告。テレメトリが黙って空になるのを防ぐ 📌 修正31件、詰まり系と壊れた履歴の救済が中心 ・復元されたメモリファイルの「経過時間」の注記がcompactionや再開のたびに書き換わり、プロンプトキャッシュを外していた ・Grep／Glob／@ファイル候補が20MBの出力上限を超える検索で固まる、またはメモリ切れ。ripgrepが警告の洪水のあと「no matches」と誤報する件も ・Readツールがメモリ圧で大きいファイルの一部をデコードできないとき、エラーを返さず固まっていた ・保存済みトランスクリプトに壊れたエントリがあると `--resume` `--continue` や起動そのものが失敗していた件が3本まとめて ・APIのエラー応答を書き換えるネットワークゲートウェイ配下で、毎ターン「API Error: 400」になっていた ・Linuxのサンドボックス内Bashで、zshだと失敗したコマンドが終了コード0を返していた ・フォーク／背景セッションの `/rewind` が、バックアップを写し切れていないとゼロ埋めや途中切れのファイルを復元していた ・応答に `</ccmemory>` みたいな閉じタグが混ざる件 ・プラグインのログや `claude plugin marketplace list` に、git／ssh／マーケットプレイスURLに入れたパスワードやトークンが出ていた ・`/update-config` が `Write(path)` の権限ルールを書いていた。ファイル権限チェックは `Edit(path)` を見る ・`hooks/` `config/` という名前のプロジェクト配下ディレクトリに、サンドボックスBashが書けなかった 📌 変更で挙動が変わるところ ・npmソースのプラグインは `npm pack --ignore-scripts` で取得して整合性検証。パッケージのインストールスクリプトはもう走らない ・貼り付け／添付した画像は、権限プロンプトなしでClaudeが開ける場所に保存される（DesktopとVS Codeも） ・Chrome拡張のautoモードは、分類器が通した呼び出しでは拡張側のサイト別チェックを飛ばす。リダイレクト後の `browser_batch` が「Permission denied」になっていた件の対処 ・`--system-prompt` に `__SYSTEM_PROMPT_DYNAMIC_BOUNDARY__` の行があれば、その上の文面はグローバルにキャッシュ ・プラン使用量の読み取りは、同じ端末のエディタ窓や非対話セッションで1分以内なら共有 📌 VSCode 27件、diffの変更ごとに採否が決められる ・提案変更のdiffタブに、変更ひとつずつ「承認／却下」ボタン ・Memoryダイアログで保存済みメモリの閲覧・編集・削除 ・画像を添付するだけで、文字なしでも送れる ・返答が流れている最中にスクロールを上げても下に引き戻されなくなり、`claudeCode.scrollToBottomOnSend` で送信時のジャンプも切れる ・応答中に打ったスラッシュコマンドが本文としてモデルに送られていた件、実行中セッションの名前変更が元に戻る件（2.1.269からの退行）を修正 📌 Web／Claude Tag／Code Review ・許可ドメインのリストが長すぎる環境が、保存は通るのに毎回セッション開始で落ちていた。保存時点で弾いて削る量も教える ・Claude TagにCloudWatch／SNS／Google Cloud Monitoring・Loggingのプリセット、DatadogのUS3・AP1・AP2・US1-FEDサイトも ・他のbotが@Claudeしたとき二重に返事することがあった件 ・レビューが100回を超えたPRが、ベースからのマージのたびに全件再レビューになっていた 手元も朝の時点で275に上がってました。96件のうち修正31、そのうち「壊れた履歴からの復帰」だけで4本。長く育てたセッションを大事にする方向で、ありがたいです☺️ https://code.claude.com/docs/en/changelog #ClaudeCode #AI