Community posts

「Claudeのmodの使い方よくわからないな〜」っていう人、まずこれやってみて！ /plugin enable cc-plugin-you-should-know@builtin と打つと、別エージェントが作業を横で見て、自分や Claude が見落としそうなことを知らせてくれます！ これは公式のbuilt-inスキルだから、安心です((´∀`*))

You can now vibe-code Claude Code itself. Anthropic just shipped Mods: custom UI and code that changes how the agent handles actions. For app builders: put live build status inside Claude and make it ask before touching production config. Ask Claude to build the mod. It writes the code, installs it, and hot reloads it in your session. Start with the annoying instruction you repeat on every project. Turn it into a mod that runs automatically.

Today in AI coding: • Claude Code adds mods that rewrite its interface • Pi 1.0 ships alongside Pi Durable • OpenAI resets ChatGPT limits after Sol's slow start • GitHub's Copilot can now click around your desktop https://www.thenewway.ai/claude-code-adds-mods-plugins-that-reach-into-its-interface/

Mods are here, so I made Claude Code a tiny coworker 👾 Clawd lives next to the thinking line and acts out what Claude is doing: cat README.md → a cat walks in git blame → he points at you git push --force → 💥 74 acts, free to install 👇 https://github.com/raresmun/claude-mods

Claude Code plugin "md-prompt" is out! It paints Markdown onto the prompt box as you type. Fenced code becomes a syntax-highlighted card before you even close the fence. /plugin marketplace add nogu66/md-prompt /plugin install md-prompt@nogu66 Check it out!

🚨 Anthropic تنشر الدليل الكامل لبناء الـ Mods لـ Claude Code.. ​نشرت Anthropic دليلها الرسمي لإطلاق الـ Mods، لكن السحر الحقيقي هو استخدام Opus 5.5 لفحص أرشيف جلساتك السابقة واستخراج الـ Mods الأكثر أهمية لإسطبلك البرمجي بناءً على أسلوب عملك! ​📌 كيف تُغيّر الـ Mods تجربة كودك؟ الـ Mod هو ميزة صغيرة تُضيفها لـ Claude Code لتُظهر لك المعلومات الحيوية مباشرة على الشاشة دون أن تضطر لسؤاله كل مرة: حجم الـ Context المتبقي، ما هي الملفات المحذوفة قريباً، والتعديلات الفورية! بدلاً من نسخ الأمثلة جاهزة، اجعل النموذج يكتشف احتياجك الخاص بنفسه. ​🔥 البرومبت السحري لإرساله فوراً لـ Opus 5.5 👇

Claude Code users can now ask it to redesign itself. Mods can rewrite prompts, block tool calls, and redraw the interface, all from a plugin. https://thenewstack.io/anthropic-claude-code-mods-plugins/?taid=6abfc706adc2d2000147994b&utm_campaign=trueanthem&utm_medium=social&utm_source=twitter

bookmark and use this prompt to get a list of 10 useful Claude Code mods to improve your productivity based on how you use Claude ----- look through my past transcripts and your memory on how i use claude code in the desktop app. research the new claude mods feature that recently released. compose a list of 10 mods that would improve my productivity. -----

Addicted to Claude Mods now

Claude Codeの新機能、Modで早速作ってみた。 1つ目：横にファイルブラウザと、クリックでパスをコピーするMOD(相対パスモードは@つき。絶対パスモードは＠なし) 2つ目：よく使うコマンドをワンクリックで実行できるように下にボタン配置

Just built myself a /deploy plugin showing in a quick glance my cloudfare deployments current state. Claude Mod is something that everyone should try. Just ask claude code to read your chats and analyze and suggest what will be the best plugin/mods to make based on my workflow.

@ClaudeDevs I'll try the built-in You should know mod on a real overnight session first. Side agent catching what I rubber-stamp is the scar I actually have. writing my own can wait until I've run claude plugin validate on something tiny.

I did as they told me to; I used the mod feature and modded Claude Code. Told it to make its own theme and add a widget where it shows its emotion. I actually did not expect it to show emotions, so i was testing it, but Opus 5.5 noticed it and added it... now i am even more confused.

claude code sacó mods un mod es un código javascript que corre adentro de claude y puede mirar, cambiar o frenar cada agente trae uno que se llama "you should know" que es un segundo agente que no hace nada, solo mira la sesion está bueno para supervisar corridas largas de trabajo hay que tener cuidado con los permisos porque corre fuera de sandbox y ve todo, tus prompts, puede frenar agentes, ve archivos y claves, etc pero esta bueno porque te puede dejar un resumen de todo lo que paso y evitar que los agentes se paren o hagan cosas que no deben voy a ver de usarlo en la próxima corrida larga https://code.claude.com/docs/en/plugins/mods/overview

DAILY AI BRIEF 📋 Oct 2 ANTHROPIC > Claude Code mods are live. Small TypeScript functions that rewrite prompts, block or retry tool calls, gate permissions, or draw UI. > Ship inside plugins via /plugin. CLI and desktop. Some built-ins like /diff are mods you can disable or replace. > Not sandboxed. Same machine access as Claude Code. Trust the source like any local install. https://claude.com/blog/claude-code-mods GITHUB > Copilot computer use in public preview. CLI and Copilot app on macOS and Windows. > Click, type, scroll, drag across desktop apps. Aimed at GUI only and legacy tools with no API, CLI, or MCP. > Off by default. /computer on. Approvals before control. Orgs can disable. https://github.blog/changelog/2026-10-01-github-copilot-can-now-interact-with-desktop-apps/ GOOGLE > Mantis is an open skills pack for coding agents: threat model, scan, filter false positives, sandbox reproduce, patch, report. > Reproduction runs generated code. Isolate it. Not a supported Google product. https://github.com/google/mantis Hooks that rewrite tool calls beat another chat wrapper. Desktop control without a kill switch is how demos become incidents.

Claude Code can now build and flash your ESPHome devices, and you can watch it happen 🔧 Free with the Vome app for Home Assistant, no sign-up, no ports opened: https://github.com/Vortitron/home-assistant-mcp/tree/main/claude-plugin/vome-esphome

I built a position tracker in Claude Code with Claude’s new Mods. No more tab-switching to check whether you’ve been rugged. See your live positions while you’re building your next side hustle. You can also check a token’s price, transactions, and chart directly inside Claude Code, so you never miss a move. The gap between traders and builders is converging to zero. The repo for the Claude Mod is here: https://github.com/nathanliow/meme-watch

Anthropic just published its 𝗳𝘂𝗹𝗹 𝗴𝗲𝘁𝘁𝗶𝗻𝗴-𝘀𝘁𝗮𝗿𝘁𝗲𝗱 𝗴𝘂𝗶𝗱𝗲 for Claude Code mods A mod is a small feature you add to Claude Code. The guide's three mods take the things you'd normally stop to ask it and 𝗽𝘂𝘁 𝘁𝗵𝗲𝗺 𝗿𝗶𝗴𝗵𝘁 𝗼𝗻 𝘀𝗰𝗿𝗲𝗲𝗻: how full the context is, what this command is about to delete, and what it just changed. Keep mods to yourself, or submit them to the Claude directory if others might find them useful. Mods run with the same access to your machine as Claude Code itself, so make sure to only install mods from sources you trust.

This is a pretty interesting step for Claude Code. Instead of treating the tool as something you just use as, you can now change how it behaves, customize the interface, and even build your own features on top of it. That makes Claude Code feel less like a fixed coding assistant and more like something developers can actually shape around their own workflow. And the interesting part is that you can build these mods with just a few lines of TypeScript or let Claude build them for you. AI coding tools are slowly becoming platforms you can customize, not just tools you use.

claude code got mods yesterday. so i put codex inside it talk to codex out loud, it hands the work to claude codex makes images on your chatgpt plan claude hits its limit? codex picks up the same chat it's called nightshift. open source, link below

October 1st, Claude Code shipped the ability to rewrite its own behavior from the outside. Mods are plain TypeScript functions hooked into the events Claude Code already emits: a tool call, a permission prompt, a UI render. A mod can run before, after, or instead of any of them. Anthropic's own proof of commitment is that /diff, a feature that used to be hardcoded, now ships as a mod you can disable or replace. This is a different bet than the rest of the editor market made. Cursor inherited VS Code's extension marketplace wholesale, thousands of plugins, broad surface area, a decade of precedent. Claude Code went the other way: a small, surgical hook API that intercepts specific moments in an agent loop rather than extending an IDE shell. It is closer to a kernel module than a browser extension. That buys power. A mod can block a risky shell command before it runs, redact a secret from an output, or swap a built-in feature entirely. It also buys risk: mods run with full machine access and nothing sandboxes them, so the entire security model shifts to trusting whoever published the plugin. That tradeoff is the real story here, not the feature list. Every agent platform is now choosing between a wide, familiar extension surface and a narrow, deep one that lets you rewire the agent's actual decision loop. Claude Code just picked the second, for a terminal tool that already runs with your repo's full permissions. Worth watching whether Codex or Windsurf answer with the same hook model or stay with the marketplace approach. https://x.com/ClaudeDevs/status/2105721434807083061

Learn to build a custom harness, folks. Mods in Claude Code, DeepSeek Harness, and now Pi Durable. What do these all have in common? Increased extensibility. Builders need coding agents and harnesses to be more malleable. Say you are building a simple agentic app for your users, a personalized headless experience, a software factory, or a domain-specific harness; you will quickly find that the generic harness won't cut it. Sometimes you need a different behavior in the UI. Sometimes you just need to easily add a new feature/plugin. Or sometimes you just want to build a completely new experience, like multiplayer support. You can build it from scratch, as I recommend. I mean, it's not that hard. But if you want to hit the ground running, where do you go? Pi Durable and DeepSeek Harness seem to want to solve that problem. Users want more customization in their harnesses. And those builders who understand how to build and maintain one will win big time from this next phase of more malleable harnesses. The battle of the best harness will be very different from the first wave of generic harnesses.

Eu parei de instalar mods na mão. Comecei esse experimento com Control Resonant. O máximo que eu faço é baixar o arquivo, aponto a pasta pro Claude junto com a pasta do jogo e mando ele se virar. Tive 0 problemas. Instalou melhor e de forma mais organizada do que eu faria.

Claude Code mods can rewrite prompts and change tool calls. Before installing one for a nicer interface, check what else it can change. Review its code and setup guidance. A text scan cannot certify executable code. https://sunglasses.dev/blog/skill-store-is-new-package-registry #AgenticAISecurity #AISecurity #cybersecuritytips #Sunglassesdev #PromptInjection #ClaudeAI

SOMEONE GAVE CLAUDE CODE THE KEYS TO ANY PC GAME universal-modder: you point claude code at a game you own and it figures out the engine, writes the mod, makes the art and sounds, then tests it in the running game works with unity, minecraft and 10 other engines, and backs up your saves before touching anything. art and audio need a fal key their example prompt puts real minecraft inside GTA V story mode 1.8k stars in 2 days

We still cannot change the cursor shape using claude mods so vim cursor shape change is still not doable 😔 @ClaudeDevs

Claude Code’a yeni gelen Mods özelliğini görünce tabii ki ilk yaptığım şey verimlilik aracı olmadı terminalde claude çalışırken, Minecraft temalı Clawd ile oynayabileceğim küçük bir oyun yaptım.

My first mod: make Claude same as Codex 100%

Claude Code 推出了新功能 Mods，来看看它能做什么、怎么用，以及和其他几家（Codex、DeepSeek Harness 、LangChain 中间件）的实现，有什么不同？ https://x.com/i/article/2105988602106884096

🔥 Claude Mods وصلت مش plugin برا — كود يشتغل جوّا Claude Code نفسه. يعدّل tool calls · يرسم UI · يقدر يوافق قرارات لوحده. خطر: صلاحيات جهازك كاملة. ثبّت بس من مصدر تثق فيه. https://claude.dev/blog/getting-started-with-claude-code-mods/