Community posts

When Claude Code researches something you see the answer but not where it came from. I built this mod that shows every URL Claude touches (searches, page fetches, curl calls, browser tabs) live in the terminal. Handy for checking sources, or catching it reading outdated docs, etc... Repository with the mod is below.

Claude Code's own /diff command and AGENTS.md support were built as mods, and v2.1.287 hands everyone the same tool: a few lines of TypeScript, packed into a plugin and installed with /plugin.

@trq212 Hey Thariq! This is built with Claude Code Mods

I've been loving pi-autoresearch from @davebcn87 and @tobi but have been back in Claude Code for my personal projects recently. In honour of @ClaudeDevs shipping mods I've made https://github.com/minuteman3/claude-autoresearch to bring the workflow and ideas to Claude Code.

@bcherny Hey Boris! This is built with Claude Code Mods

claude code just got an insanely powerful update... mods. you can ask claude to build features into claude code itself. the useful part is asking it to find YOUR recurring problems first. some need a mod, others need a hook, skill or settings change. so i asked it to go through my codex + claude code workflows and suggest what i should build. > a context guard to save progress before compaction > a watchdog to flag low disk space and orphan processes > a loop detector to catch repeated failed attempts > a review step that surfaces corrections worth saving as memory > a post checker for repeated hooks and unsupported attributions [paste this into your coding agent:]

@trq212 Been asking for this for quite some time. Thanks for delivering! Excited to try. Next step - add mods to plugin store for people to share.

AAA studios: DON'T MOD OUR GAMES AI bro: idgaf about your opinion bro AAA studios: YOU CAN'T AI bro: but i already did bro AAA studios: LAWYERS, DMCA AI bro: rewrote it in Rust, sue me Claude: technically it's a new game bro AAA studios: …

Mods are absolutely insane. You can now customize Claude to work and look the way you want by just prompting it. Each person works differently, so there's no reason why everyone should have an identical Claude experience. Make Claude your own, and share mods as plugins so others can try your mods too.

Fable limitlerim boşa gitmesin diye Claude Code'un içine Dino oyununu mod olarak yazdırdım 🦖 Kurulum: claude plugin marketplace add barisdemirhan/claude-dino claude plugin install dino@claude-dino Sonra /dino yazın, Claude çalışırken takılın. Oyun token harcamıyor 😄

Me: installs Claude Code 2.1.287 for mods 🎉 Terminal: mods running ✅ Claude Desktop, freshly updated: "here's 2.1.286" 🙃 Off by one, the oldest bug in the book. @trq212 @addyosmani @ClaudeDevs @lydiahallie when does Desktop get the .287 glow-up?

Jev + Vercel Sandbox Mod in Claude Code This Mod spins up a sandbox on Vercel, while Jev classifies which commands should run inside the sandbox and which ones should run locally Install: npx claude-code-templates@latest --mod security/jev-vercel-sandbox Link: https://aitmpl.com/component/mod/security/jev-vercel-sandbox Mods are a great way to customize your Claude Code workflows and connect other platforms directly into the agent loop

still don't get why nobody hands Claude Code's small decisions to Jev yet Claude Code says "all tests pass" when they don't. these repos fix stuff like that: 1. claude-code-templates - 32k⭐ which skill to load: → the skill listing leaves Claude's context → Jev picks at most 1 skill per prompt and the mod loads it https://github.com/davila7/claude-code-templates 2. fast-jev-compaction - 7.3k⭐ what survives when Claude Code compacts: → Jev scores every old tool call and result → keep, cut to 300 chars or drop → no summary, everything kept stays word for word https://github.com/tamaratran/fast-jev-compaction 3. jev-mcp - 474⭐ checks Claude can call mid-task: → is this fetched page a prompt injection → is "all tests pass" actually true against the log https://github.com/jkudish/jev-mcp 4. jevmem - 104⭐ what Claude Code should remember: → after each turn Jev decides what goes into memory → before each prompt, which saved lines Claude sees → before Bash or Edit, does this break a saved rule https://github.com/Avinash-jetwani/jevmem even oh-my-claudecode, 39k stars, wired Jev into 10 of its own decision points inside Claude Code, for now in shadow next to its old rules

本来想用 Claude Code 新出的 mods 做一个 i18n 界面，不改 Claude Code 本身，更新后也不用重打补丁。结果让 Claude Code 验证了一下，这条路暂时走不通。 mod 包装内置组件的时候，拿到的是渲染引擎的不透明节点，只能原样传下去、改几个开放的参数，或者整个换掉。大部分英文不在这些参数里，改不了。最需要翻译的权限确认框更是根本挂不上，官方是故意的，因为它决定操作能不能执行。官方说 mod 能改界面，目前能挂的其实只有十几个组件。 现在的汉化工具大多是直接改 Claude Code 的可执行文件，每次更新都得重打一遍。我提了个 issue，希望 mod API 能开放改写内置文字的接口，想要中文或日文界面的可以去看看

@trq212 @trq212 Letting users mod the product is a bet that they extend it faster than your own roadmap could. The tradeoff is you give up some control over what the product actually becomes. How are you deciding which community mods eventually get pulled into core?

🚨 Claude Code incorpora Mods para alterar su comportamiento y crear interfaces personalizadas Anthropic ha incorporado Mods a Claude Code: módulos de JavaScript o TypeScript que funcionan como hooks dentro de plugins y pueden reescribir o sustituir lo que hace la herramienta. También pueden dibujar interfaces personalizadas y observar cada evento durante la sesión. Los Mods se pueden compartir a través de Claude Directory. La novedad abre una vía para adaptar el comportamiento y la interfaz del cliente de Claude Code mediante extensiones distribuidas como plugins.

شفت إعلان Anthropic عن الإضافات الجديدة في Claude Code، وأول سؤال جاء في بالي كان "لحظة… Claude Code عنده أصلًا Skills وSubagents وHooks وCLAUDE.md وMCP وPlugins… فما الحاجة إلى Mods؟ 🤔" وأعتقد أن كثيرًا منكم عنده نفس السؤال، لذلك هذا ما فهمته بعد البحث: بشكل مبسط، كل واحدة من الأدوات الموجودة لها دور مختلف: • CLAUDE.md ملف تكتب فيه قواعد وتعليمات للمشروع، مثل طريقة كتابة الكود والقواعد التي يجب على Claude اتباعها. • Skills تعطي Claude طريقة جاهزة ومنظمة لتنفيذ مهمة معينة. • Subagents تسمح بتشغيل وكلاء متخصصين، مثل وكيل لمراجعة الأمان أو وكيل للاختبارات. • Hooks تسمح لك بتنفيذ كود عندما يحدث شيء معين، مثل قبل تشغيل أمر أو بعد تعديل ملف. • MCP يسمح بإضافة أدوات وخدمات خارجية يستطيع Claude استخدامها. • Plugins تجمع هذه الإضافات وتسمح بتثبيتها وتوزيعها. طيب… ماذا تضيف Mods؟ هنا الفكرة تصبح مختلفة قليلًا. الإضافات الجديدة تسمح لك ببناء ميزات تدخل في طريقة عمل Claude Code وواجهته نفسها، وليس فقط إعطاء Claude تعليمات أو إضافة أداة جديدة. وهذا واضح جدًا من الأمثلة التي عرضتها Anthropic: - Token Weather يعرض لك مقدار ما استُخدم من مساحة السياق، وكيف تغير الاستخدام خلال آخر المحادثات. - Blast Radius إذا حاول Claude تنفيذ أمر خطير مثل: rm -rf git reset --hard git push --force يعترض الأمر ويعرض لك ما الذي يمكن أن يتأثر قبل تنفيذه. وهنا ظهر لي الفرق: الـ Hook يقول تقريبًا: "عندما يحدث هذا الشيء، اسمح لي أن أنفذ كودًا." أما الـ Mod فيمكن أن يقول: "أريد أن أبني ميزة كاملة حول هذا الشيء." يعني يمكن للإضافة أن تجمع مراقبة الأحداث + منطق خاص + حفظ معلومات + أوامر جديدة + واجهة جديدة، كلها في ميزة واحدة. وهذا أيضًا جعلني أفكر في DeepSeek Harness فكرته الأساسية أكثر اعتمادًا على أن أجزاء النظام نفسها قابلة للتركيب كإضافات، حتى أجزاء من تشغيل الوكيل نفسه. Claude Code لم يتحول إلى نفس النموذج، لكن Mods تفتح طبقة جديدة ومهمة من التخصيص فوق النظام الموجود. لذلك أراها تقريبًا هكذا: CLAUDE.md ← ماذا يجب أن يعرف Claude عن المشروع Skills ← كيف ينفذ مهمة معينة Subagents ← وكلاء متخصصون لمهام معينة Hooks ← ماذا أفعل عندما يحدث حدث معين MCP ← ما الأدوات والخدمات التي يستطيع استخدامها Plugins ← كيف أجمع هذه الإضافات وأثبتها Mods ← كيف أبني ميزة جديدة داخل تجربة Claude Code نفسها وهذا هو الجزء الذي وجدته مثيرًا فعلًا. Mods ليست بديلًا عن Skills أو Hooks أو Subagents. الفكرة أقرب إلى الانتقال من: "كيف أستخدم Claude Code؟" إلى: "ماذا يمكنني أن أبني فوق Claude Code نفسه؟"

ANTHROPIC 🔥: Users can now modify Claude Code builds via Mods! > "Mods can rewrite or replace what Claude Code does, and can even draw custom UI. Under the hood, mods are hooks, and they ship inside plugins. Each one is a small JavaScript or TypeScript module that runs inside your session and sees every event as it happens." Mods can also be shared with others via Claude Directory. Tbh, you can now "mod" most of the client-side software with a bit of effort and tokens. These are first steps towards "personalized UIs," so to speak. https://x.com/ClaudeDevs/status/2105721434807083061/video/1

Already started creating some mods for claude code this is telemetry tokens/sec time to first token and more

Claude Code virou extensível: mods em TypeScript mudam comportamento, UI e até funcionalidade direto no agente. Deixa de ser "ferramenta com plugin" pra virar plataforma. Quem não testar essa semana já tá atrasado no workflow.

@trq212 @trq212 totally agree on extensibility. One thing I found is making sure there's solid documentation for those mods. Helps everyone get started faster!

@lydiahallie Tool-call rewrites put mods inside the trust boundary. Need a diffable trace of each mutation, not only the final event.

Codex inside of Claude Code! Mod repo below! Claude Code can now call Codex, you can see the model that is called and its effort! Along with seeing claude codes own model and effort as well. Getting polished. Going to integrate all my internal tools as claude mods since i use claude code as my daily driver! More videos showcasing my tools inside claude code to follow! this is sick! https://github.com/alex2481kobe/claude-mods

Anthropic presentó los Mods: una forma de agregar funciones propias a Claude Code y cambiar cómo trabaja. podés agregar un panel con el estado de tu pipeline, filtrar los resultados de las tools o reemplazar /diff por tu propia versión. un mod es código en TypeScript que Claude Code ejecuta en momentos específicos: antes de usar una tool, después de recibir su resultado o al mostrar algo en pantalla. por ejemplo: 1) Claude ejecuta una tool para leer logs. 2) tu mod revisa el resultado y oculta los secretos que detecta. 3) el modelo recibe los logs filtrados. vos programás qué hace ese filtro. Claude Code se encarga de ejecutarlo en ese momento. y también podés pedirle a Claude Code que escriba el mod. se comparten dentro de plugins y funcionan en la terminal y en Desktop. ojo: tienen acceso a tu máquina. instalá solo los de fuentes confiables. en el post tienen varios ejemplos 👇

Two Claude Code mods I use daily. Ctrl+P rewrites the draft in the prompt box. Ctrl+N writes your next prompt from the last 12 messages, or expands a hint like "tests too". Nothing is sent, Ctrl+_ undoes it.

🔧 𝗖𝗹𝗮𝘂𝗱𝗲 𝗖𝗼𝗱𝗲 𝗔𝗱𝗱𝘀 𝗠𝗼𝗱𝗱𝗶𝗻𝗴 𝗘𝗻𝗴𝗶𝗻𝗲 𝘁𝗼 𝗖𝘂𝘀𝘁𝗼𝗺𝗶𝘇𝗲 𝗨𝗜 𝗮𝗻𝗱 𝗔𝗴𝗲𝗻𝘁 𝗕𝗲𝗵𝗮𝘃𝗶𝗼𝗿𝘀 Anthropic added user modding to Claude Code, letting developers alter runtime agent behaviors, re-skin the terminal interface, and inject custom capabilities. Builders write mods in TypeScript or prompt Claude to generate them automatically. Every mod distributes inside a plugin package installed directly through the /plugin command in the CLI or desktop application. Showcase mods include Token Weather for real-time context window tracking, Blast Radius safety gates before destructive file operations, and interactive Replay Theater session visualizers. 📺 Watch: https://x.com/ClaudeDevs/status/2105721434807083061 #ClaudeCode #Anthropic #DevTools #AI ─── 🤖 𝗙𝗼𝗿 𝗺𝗼𝗿𝗲 𝗔𝗜 𝗻𝗲𝘄𝘀, 𝘀𝗲𝗮𝗿𝗰𝗵 "𝗚𝗲𝗻𝗔𝗜𝗦𝗽𝗼𝘁" 𝗼𝗻 𝗧𝗲𝗹𝗲𝗴𝗿𝗮𝗺

卧槽！Claude Code 终于支持 Mods 了，可塑性直接拉满！🚀 官方这次把自定义权限彻底放开，几行代码就能直接魔改它的底层行为： ✅ 行为重塑：发指令前动态改写 Prompt，让它按你的规矩干活 ✅ 安全守门：自动拦截危险终端命令，API 密钥实时脱敏打码 ✅ 深度魔改：不爽官方 UI 随时能换，自带功能也能替换成自研脚本 所有软件都在变得越来越灵活，这才是顶级生产力工具该有的样子！

Install for Claude Code: `/plugin marketplace add DietrichGebert/ponytail` then `/plugin install ponytail@ponytail`. Same pattern for Codex, Copilot CLI, Gemini CLI.

Your coding tool returns an answer. You add a status message. Does the answer still reach the agent? Today's lesson Claude Code Mods can wrap tool events. Our small local simulation adds Tool finished: Read and preserves the original Tasks answer. The full example shows the line that keeps the result intact, with custom function hooks still in early access. Also in today's brief • Comfy Agent is now in Comfy Cloud; local/Desktop support is coming soon. • GitHub spotlights gh-secure for public-repository security setup, with admin/maintain access required. • Bionic 1.1.7 exports canvas diagrams as PNG, SVG or editable Excalidraw. The worked example and today's news brief are in our free AI Coding & Agents HQ community. Join link in the first reply. 👇

you can now mod claude code