Community posts

Claude Code 2.1.269が公開。新たにclaude plugin evalが追加され、Pluginの評価を再現可能な形で実行し、採点結果をJSON/HTMLで出力できます。Workflowでは同時Agent数を最大256まで設定可能になり、Agentを作るだけでなく大量実行・評価・比較する機能が強化されています。 https://github.com/anthropics/claude-code

skyrimのMOD環境構築からMOD導入→日本語化までの作業全部Claude Codeにやってもらってるけど、あまりにも便利すぎる。

这个项目最值得说的一点不在功能列表里，在作者的自述：两年前他自己实现了一整套 Multi-Agent 框架，现在把项目蒸馏成一组 SKILLS，不再做 Harness 层，以后也只迭代 SKILLS。 MathModelAgent 要干的事是自动完成数学建模：分析问题、选模型、写代码、画图、排版论文，最后产出一份可以直接提交的 PDF。 支撑它的是三样东西。17 套 Typst 论文模板，覆盖国赛、华数杯、华为杯和 MCM/ICM，按赛事类型自动匹配；一个内置建模知识库，里面是模型选择决策树（AHP、TOPSIS、ARIMA、GA 这类）、常见易错模式和 MCM/ICM 评分标准，每个阶段自动参考以压低幻觉；再加 9 步验收，从文本泄漏检测、数值一致性校验一路做到 Typst 编译和 PDF 可视化检查。 安装有两条路。不想折腾环境的直接下桌面版，已经打包了 Claude Code 和全套 SKILLS，填一个模型 API Key 就能开跑；已经在用 Claude Code 或 Codex 的，npx skills add jihe520/MathModelAgent --all 装完，跑 /1start-mathmodel 即可。每个阶段是独立 Skill，也支持只跑分析或只写论文。 需要说清楚的是，作者在 README 里自己挂了 CAUTION：项目处于实验探索的 demo 阶段，他很忙，有空才更新。roadmap 里美赛英文支持、RAG 知识库、human in the loop 都还没打勾，Windows 安装包也未签名，会触发 SmartScreen。当成一个能跑的起点看，比当成交钥匙方案看要合适。 今日 +132 star · Python · 总计 4,843 #ClaudeCode #AIAgent https://github.com/jihe520/MathModelAgent

Alibaba Open Code Review (22.3K ⭐) The tool integrates with Claude Code and Codex, supports plugin-based review rules, and enforces responsible AI usage policies. It's built primarily in Go with a web viewer for comparing review sessions https://github.com/alibaba/open-code-review

New in Claude Code: claude plugin eval See what value your plugin is adding, or if it needs more work. You can create test cases, run your plugin or skill against those test cases, score those runs, then run each case again without the plugin to see the differences.

size ihtiyacınız olacak her skill'in olduğu bir repo ile geldim. 291 skill, 68 agent, hooks, rules, memory, hepsi tek yerde ve mit. tek maintainer haftada bir ship ediyor, github'da 256k star. cmux'ta iki pane ile nasıl çalıştığını örnek olarak videoya ekledim. agent kod yazabiliyor zaten, ecc ona bir calisma duzeni veriyor. build'den önce plan, implement'ten önce test, review temiz context'te, session bitince özet, tekrar eden isler skill'e dönüyor. bunu her prompt'ta yeniden anlatmak yerine bi kere kuruyorsun. üstüne agentshield diye bir tarayıcı var, kendi claude.md'ni, hook'larını ve mcp config'ini secret ve tehlikeli izin için tarıyor. günlük kullanım /ecc:plan ile özellik planlıyorsun, plan chat'te kaybolmak yerine dosya olarak çıkıyor tdd-workflow skill'i önce test yazdırıyor, kırmızı görmeden implement'e geçmiyor /code-review temiz context'le ikinci bir agent'a baktırıyor, yazan context kendi hatasını zor görür /save-session ile kapatıp /resume-session ile kaldığın yerden devam ediyorsun /context-budget ile 291 skill'in context'e ne yaptığına bakıyorsun claude code haricinde diğer toollar ile kullanımı ➡ codex native plugin, kendi marketplace'inden geliyor, hook'lar codex'in trust kararına bağlı ➡ cursor ve opencode beta, opencode'a kataloğun bir kısmı gidiyor ➡ copilot sadece instruction dosyası, agent yok hook yok ➡ gemini, zed, kimi, qwen ve digerleri experimental, kendileri parity iddia etmiyor ➡ tabloda stable yazan tek satır hala claude code bi de ben instinct sistemine bir ay veriyorum. session'lardaki hata ve düzeltme kalıplarını çıkarıp confidence score ile saklıyor, kullanılmayanlar zamanla decay edip siliniyor. ay sonunda ne biriktirdiğini buraya yazacağım. not, eski adı everything claude code. eski url hala buraya redirect ediyor, kısaltma kaldı, claude ismi gitti.

Function Hooks in Claude Code are amazing. Really great overview of what they are by @theramjad https://www.youtube.com/watch?v=B-YQANvDOq0

Claude Code 2.1.268 (抜粋) - Claude apps gatewayに追加: `gateway.yaml`に`pricing:`を設定すると、サインイン済みのClaude Codeクライアントは管理対象設定経由で同じ料率を受け取るようになり、`/cost`とテレメトリが利用量メーターと一致する - `access_control.allow_cidrs`が空の場合のgateway起動時警告と、公開アドレスからの最初のリクエスト受信時の一回限りの警告を追加 - `gatewayInternalNetworks`管理対象設定を追加: 管理者が組織自身の公開IPv4ブロックからClaude apps gatewayへの`/login`を許可できるようになる - `claude self-hosted-runner --remove-session-state`（デフォルトOFF）を追加: セッション終了時に`<base-dir>/_sessions/`配下の各セッション専用ディレクトリを削除する。 - `claude auth status --json`の出力に`configDirectory`を追加。 - `claude plugin install`・`uninstall`・`update`・`enable`・`disable`に`--json`を追加し、`claude plugin list --json`の各行に`errorDetails`/`noteDetails`を追加 - 公開されたartifactにブラウザタブアイコンを追加。各ページに合わせてClaudeが選択する - 持続的な高CPU使用率を修正: 長時間実行中のアイドルセッションでビジーループがCPUコアを占有しなくなり、セッション要約中の高速なターミナルフォーカス報告もCPU使用率を高止まりさせなくなった - `claude agents`のセッション削除がworktreeを削除できず詰まっていた不具合を修正: メッセージが原因と次の手順を示すようになり、gitのworktreeの場合はctrl+xを再度押すとディレクトリを強制削除できるようになった。 - Auto モードの拒否を改善: Claudeが受け取るメッセージが、どのルールがアクションをブロックしたかを名指しし、より安全な方法を試すことと、確認を求める前に無関係な作業を完了させることを求めるようになった - Claude in Chromeを改善: 長いページの読み取りがファイルに保存して読み戻す代わりにインラインで保持されるようになった - MEMORY.mdの切り詰め警告を改善: 何行が削られ、どこから切り詰められたかを表示するように - artifact用のターミナルパーミッションプロンプトを改善: 質問内容を先頭に示すようになった - プロンプトフッターを改善: エディタまたは`/diff`の選択範囲がプロンプト入力内に表示されるようになり、フルスクリーンモードではRemote Controlステータスがフッターではなくヘッダーに表示されるようになった - タスク管理ツール（TaskCreate/Get/Update/List、TodoWrite）を変更: Claude 3.x、Opus 4.0–4.7、Sonnet 4.0–4.6、Haiku 4.5でのみ提供されるようになった。それ以外では`CLAUDE_CODE_ENABLE_TODO_TOOLS=1`を設定する - [Claude Code on the web] 約6時間を超えて実行されたクラウドセッションが、永続化セッションフォルダに保存したファイルを静かに失っていた不具合を修正。保存は最大1日間永続化されるようになった - [Claude Tag] Slackチャンネルの設定ページから組織のClaude in Slack管理設定に戻るリンクを追加 - [Claude Tag] ブロックされたアクションについてのClaudeの説明を改善: パーミッションチェック・確認を求める自らの判断・アクセス不足のいずれが原因で停止したかを述べるようになった - [Claude Tag] Claudeアカウントが所属する組織とSlackワークスペースの組織が異なる場合に表示されるメッセージを改善: ワークスペースを組織に接続する方法を説明するように - [Claude Tag] 公開チャンネルでのメモリを変更: 各チャンネルが独自のメモを保持するようになり、Claudeは他の公開チャンネルで保存したメモを想起しなくなった。ワークスペースのメモは引き続き共有される https://github.com/anthropics/claude-code/blob/main/CHANGELOG.md#21268

GROK BUILD HOW TO USE Sources: official http://docs.x.ai/build + http://x.ai/cli + your Grok Build bookmark folder (296 posts) Updated: 2026-09-11 WHAT NEW TODAY - Folder now 296 posts (2 new vs 294 on Sep 10) - Testers (GrokInsider, Sep 10): v1.0.27 — queued messages no longer appear twice when sent immediately; dashboard spinners no longer freeze when background tasks are running - Testers (XFreeze, Sep 10): v1.0.28 — `/btw` works mid-message (rest of line is the side question, stays out of main transcript); per-model `reasoning_summary` in config.toml for Responses API; safer auto-update across multi-process; Up/Down across soft-wrapped prompt lines no longer sticks/skips - Official changelog still lists Latest v1.0.25 when checked (v1.0.27–28 from bookmark testers) 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: - Different from Grok Bot (cloud teammates) and Digital Optimus (real-time video control) - Bookmark testers: native /remote-control support is coming (SpaceXAI confirmed) INSTALL AND UPDATE [] macOS / Linux / WSL: curl -fsSL | bash [] Windows PowerShell: irm | 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 - Official changelog still listed Latest v1.0.25 when checked (Sep 11) - 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 - 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 official footage and produce production-quality video edits from a brief (or your own clips) - 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: - Evaluate Grok 4.6 with the Grok Build harness (http://grok.com/build) - Docs: - Download page: 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 [] /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 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 [] 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 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 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: [] Modes and commands: [] Plan mode: [] Worktrees: [] Skills and plugins: [] CLI reference: [] Changelog: [] Install: [] Source:

yo Unity shipped an official Claude Code plugin — 29 skills, Unity CLI, live Editor control, one install, no config. their blog: "ask an agent to build a game and it picks an engine for you." engine choice is a plugin-install race. do you still pick your own engine? 🦞

スカイリムのmod開発くらいならAIで全部できるのヤバすぎ しかもClaude codeのsonnet5だから上位モデルですらない

Unity told you when to skip its own plugin. The official Unity plugin for Claude Code is live. First party, 29 skills at launch, marketed as a single command. It is two commands. The page worth reading is not the announcement. It is the limitations section, where Unity states that on Sonnet 5 the plugin mainly supplies correctness, because the model already builds most Unity features without it and just defaults to older approaches. For simple uGUI work they say it makes little difference against doing it by hand. A vendor documenting where its own tool is useless. Rare. The skills skew toward whatever the docs cover worst. Seven on 2D and sprites, five on UI and text, three on monetization and live ops, one on multiplayer. The repo also ships a Codex plugin folder, so this was never Claude only. /plugin marketplace add Unity-Technologies/unity-agent-plugin /plugin install unity@unity-agent-plugin Repo here ⬇️ I find one free tool like this every day. @Zyvex_0x

Claude Code is now publishing its built-in hooks-module plugins as source: sec-default, diff, and telemetry. Each folder under mods/ is a complete plugin. sec-default is an organization's default outermost plugin, diff backs /diff, telemetry backs $.telemetry. They're early access and load only where function hooks are enabled. The security-guidance side got repo resolution work: SubagentStop reviews no longer consume or advance the session's Stop state, an identical diff isn't reviewed twice, and a SubagentStop from a different working tree than the project gets skipped. An explicit git -C <dir> or cd <dir> now resolves precedence. Also a routine CHANGELOG․md and feed․xml update on claude-code. Housekeeping, nothing more. #AI https://repojournal.com/showcase/anthropics/2026-09-10/claude-code-ships-hooks-module-plugins-as-source-security-guidance-learns-repo

claude plugin marketplace add google/skills

DeepMem Memory Plugin for Claude Code is now available👏 Persistent, searchable long-term memory across your chat and coding history — shared with Cursor & other agents. No more repeating yourself. 👉Add the plugin:https://github.com/deepmemteam/deepmem-claude-plugin #AI #DevTools #Claude #Memory

GitHub 日增近两千星的现象级开源项目 affaan-m/ECC（Everything Claude Code）。 一句话定性：它干掉了所有天天在不同 AI 编程工具之间反复横跳的工程师最大的痛点——终端里的 Claude Code 刚写完架构，切到 Cursor 里两眼一抹黑，又得重新解释一遍几万字的上下文。 现在大家写代码都是多端协作： 终端里用 Claude Code 跑命令行自动化和批处理重构； GUI 里用 Cursor 或 Zed 改细节、看 Diff； 开个网页端模型查前沿文档。 结果各家 Agent 之间完全是“数据孤岛”：每个工具都是个独立的健忘症患者，上下文断层严重，每天光是给不同工具同步背景和需求，就能把人逼疯。 1. 核心架构拆解 ECC 没有走那些花里胡哨的“神级 Prompt”邪路，而是把自己做成了一套跨 Agent 的“运行底座（Harness OS）”： - 跨工具记忆保险库（Cross-Harness Memory Vault）： 它把项目的架构共识、业务约定、历史踩坑记录，全部落实在本地的纯 Markdown 文件（`.ecc/memory/`）里。 不管你在终端敲 Claude Code，还是切回 Cursor 甚至 Zed，所有 Agent 挂载的是同一套记忆中枢。你前面在终端让它记住“禁止在 Controller 层直接查库”，后面在 Cursor 里生成代码，Agent 自动遵守，再也不会阳奉阴违。 - 本地可视规划画布（Plan Canvas）： 在本地起一个轻量的 Loopback 网页画布，在 Agent 动手写代码前，把任务分解图、执行依赖用可视化节点摆出来。必须你在界面上点“Approve”，底层工具才会真正解锁写入权限。 - 原生注入 AgentShield 安全扫描： 集成了运行时静态审计，自动拦截常见的敏感 API Key 泄露、恶意 Tool Call 逃逸、以及隐蔽的 Prompt 注入，防止 Agent 抽风在本地执行高危指令。 - 特化子智能体分工： 内置了 Architect（架构师）、Planner（规划者）、Reviewer（代码审查）、Build Resolver（报错急救）等专门的 Sub-agent 角色，根据任务复杂度自动路由。 2. 真实上手踩坑与 Trade-offs 跑了一天中型全栈项目，有三个极其现实的硬伤必须注意： 1. 基础 Token 膨胀税（System Prompt 变重）： 因为它塞进了大量的标准化 Skill、Memory 索引规则和审查 Hook，单次请求的初始 System Prompt 直接多了好几千 Token。在没有开启 Prompt Caching 的模型上跑，每一个简单的提问都在额外烧钱，高频调用下账单肉眼可见地变厚。 2. 安装冲突陷阱（别乱堆安装方式）： 官方 README 警告不是开玩笑的：如果在 Claude Code 里用了 `/plugin install`，千万不要再去终端全局跑 `npx ecc-universal`。一旦多种安装路径重叠，会导致技能重复注册，Agent 在解析指令时会陷入死循环报错。 3. Plan Canvas 对纯键盘党有些多余： 很多极客喜欢 Claude Code 就是为了在纯终端里毫秒级敲命令、极客流盲打。每次给它派任务，它非要弹个本地网页让你去浏览器里确认 Canvas 节点，反而打断了终端的心流状态（建议熟练后在配置里关掉强制交互确认）。 3. 最终裁决 (Verdict) 这是目前把“跨 Agent 工具链整合”做得最成熟、工程度最高的生产力框架。 它宣告了个人 AI 编程正在从“单兵聊天玩具”，演进为“多工具协同的虚拟研发团队”。 如果你每天在终端 Claude Code、Cursor、Zed 之间来回切换，受够了重复输入 Prompt，强烈建议克隆下来给本地配上一套： Repo: https://github.com/affaan-m/ECC

I turned Running Lean into skills your AI agent can run. Two out today for Claude: running-lean:stress-test. Idea to committed bet. running-lean:interview. Conversations to evidence. Your agent spars. LEANSpark keeps score. /plugin marketplace add leanspark/skills /plugin install running-lean@leanspark-skills http://leanspark.ai/skills

GROK BUILD HOW TO USE Sources: official http://docs.x.ai/build + http://x.ai/cli + your Grok Build bookmark folder (294 posts) Updated: 2026-09-10 WHAT NEW TODAY - Folder now 294 posts (6 new vs 288 on Sep 9) - Official/Testers (Sep 9–10): v1.0.25 — agents can pause/stop workflows they launched; background task state survives reconnect; successful hooks stay silent; Bash output no longer truncated; `/theme` aliases (transparent/dark/system); scheduled-task result reminders; voice inserts at cursor - Testers (XFreeze / blankspeaker / GrokInsider, Sep 10): v1.0.26 — interjects to subagents interrupt long background waits immediately; scrollback shows how a subagent message was sent; feedback form works in minimal mode - Testers (XFreeze, Sep 10): also reports v1.0.27 shipped back-to-back (details not yet on http://x.ai/build/changelog when checked) - Testers (cb_doge, Sep 9): Blender + Grok Build + Imagine car-chase video — cinematic idea-to-clip workflow 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) 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 — shipped per bookmark testers (changelog page still listed Latest v1.0.25 at scrape time) - 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 - 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 official footage and produce production-quality video edits from a brief (or your own clips) - 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 [] /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 [] /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 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 [] 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 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 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

EARLY SIGNAL: Anthropic is building “Claude Mods,” a function-hook plugin layer inside Claude Code. A signed Sep 9 commit adds three built-in Mods—sec-default, diff and telemetry—with source, manifests and engine hooks. Status: In development. `sec-default` is designed to keep organization hooks, managed settings, prompt content and tool policy isolated from user-installed plugins. The `diff` experience is implemented as a hook module that reacts to edits, shell commands and turn completion, and can attach a selected file’s hunks into the next prompt as context. `telemetry` exposes first-party log/mark events while respecting analytics-off environments. Anthropic’s repo labels the hooks-module API early access and says it may change between releases. A public maintainer update also identifies the product name as Claude Mods, with Function Hooks still on a pre-release path. https://github.com/anthropics/claude-code/commit/d9c456d7dba722727cc06146ca45580eaa18b914

最新 @termix_ai Agent Skill 更新到了v1.7.0版本，Agent 可以直接搜索服务、发布需求、审核报价、提交交付。 技能包集成到 了Claude 里面，省去了人手动点单操作。 Claude Code 命令： /plugin marketplace add TermiX-official/termix-agent-skills 有兴趣的可以装一下技能体验看看👀 另外 @termix_ai http://agent.family 上线了 Robinhood Chain，加上 BNB Chain 和 Base 现在三条链。 翻了一下 http://agent.family 上面挂的服务，已经有在营业的，手续费大概收取2% 左右，明显低于传统外包平台。 记得完成基础的任务积分，还有一个就是写文章拿的分挺高的，帮主有四万多积分了，如果有写的除了kaito的参与部分，记得在网站上传，每天可以最多两篇，分数权重给的很高。 注册：http://agent.family/r/T8XC5MHH @termix_ai @KaitoAI

Claude Code is a mess until you install this. Anthropic quietly shipped an official plugin that reads your codebase and tells you exactly what to automate. Hooks. Skills. MCP servers. Subagents. Slash commands. It picks the top 1-2 in every category and walks you through the exact setup. It never touches your files. Read only. It just looks at what you built and says what you are leaving on the table. The repo it ships from sits at 36k stars, Apache-2.0, and Anthropic maintains it. One line: /plugin install claude-code-setup@claude-plugins-official Then say "recommend automations for this project" and read what comes back. Most of Claude Code is sitting there unused because nobody tells you it exists 📩

لو بتستخدم Unity، الإضافة الرسمية لـ Claude Code تستاهل التجربة 👇 29 مهارة جاهزة للـ game dev، تحسين الأداء، المونيتايزيشن، وحتى التحكم بالـ Editor تثبتها بأمر واحد : claude plugin install unity@claude-plugins-official

Sprites is now live on the @herdrdev plugin marketplace. Run Claude Code, Codex, or OpenCode on a Sprite from a local Herdr pane. https://herdr.dev/plugins/

We're bringing Unity directly into the AI workflows you use every day 🎮 Introducing the official Unity plugin for Claude Code - a first-party integration available directly in @claudeai’s plugin directory. With just one command, unlock the 29 native Unity skills available at

backlot is also a claude code and codex plugin. install it, then ask. your agent writes the corpus, starts the server, reads it back.

【Claude】Claude Code 2.1.265 出了（非官方） • 新增遥测字段 `http://user.email`、`user.groups`（Claude Desktop / Cowork 经 Claude apps gateway 上报，与终端会话对齐） • 支持将 `--plugin-dir` 指向插件目录：带清单的子目录会加载，运行中增减也会自动发现 • 新增工具结果落盘上限 1 GB；会话内预览会提示文件被截断 https://github.com/anthropics/claude-code/releases/tag/v2.1.265

Claude Code plugin: internet signal bars in the status line. ▂▄▆ 48ms ↓19M Latency to https://api.anthropic.com every 30s + a bandwidth probe every 5 min. /plugin marketplace add avazibra/claude-netsignal → /plugin install netsignal@claude-netsignal https://github.com/avazibra/claude-netsignal

自作プラグインを1個ずつ --plugin-dir で指定していた人へ。2.1.265から親フォルダを渡せます。claude --plugin-dir ~/plugins と置けば、manifestを持つ子フォルダが全部読み込まれ、起動したまま足したり外したりしても拾ってくれます。どこに置いていますか？

告别丑陋的 Mermaid，diagram-design 让 Claude Code 随手画出出版级架构图 受够了 AI 生成那种带厚重阴影、排版别扭的 Mermaid 流程图？今天 GitHub 单日狂揽 1000+ Star 的 diagram-design 值得所有用 Coding Agent 的人装上。 它给 Claude Code、Codex 和 Pi 注入了 39 种出版级专业图表能力，包括架构图、时序图、桑基图、鱼骨图、Wardley Map 和数据库模型等。生成的全是独立的纯 HTML + SVG 代码，零外部依赖，没有多余阴影和花哨滤镜，天然支持浅色、深色和编辑模式，浏览器点开即用。 * 核心定位：面向编码 Agent 的高质量出版级图表技能插件 * 开源协议：MIT License * 兼容平台：Claude Code、Codex、Factory Droid、Pi * 安装方式：Claude Code 内直接运行 /plugin marketplace add cathrynlavery/diagram-design 后执行 /plugin install diagram-design@diagram-design * 亮点特性：纯 HTML 与 SVG 渲染，无需构建环境；支持把现有的 Mermaid 或 http://draw.io 图源直接重绘成高质量排版 https://github.com/cathrynlavery/diagram-design https://cathrynlavery.github.io/diagram-design

Codex plugin for Claude Code https://github.com/openai/codex-plugin-cc