cad-harness

Give your AI agent CAD superpowers: real STEP and STL files, manufacturing checks and engineering drawings. 🔗 earthtojake/text-to-cad #texttocad #python #cad

Using my own stack for assembly sequence and earthtojake/text-to-cad for getting the CAD files. Thanks @earthtojake

I pointed my bug checker at itself I'm not a coder. I build things with an AI coding assistant (Claude), and a while ago I noticed something that bugged me: it would say "fixed!" and the fix would look fine, but nothing actually checked that it stayed fixed. So I made a small tool called Bite. It asks one question about every recent fix: If this change disappeared, would any test notice? It takes the change back out of a copy of the project and runs the tests. If they fail, great: a test is guarding that fix. If they still pass, nothing would warn you if the fix broke later. This week I did the obvious thing and pointed Bite at its own code. It didn't go the way I expected. Bite flagged itself, and Bite was wrong The first run flagged one of Bite's own early commits: "no test noticed it was gone." That sounded bad, so we dug in. It turned out the flag was wrong, and the reason was sneaky. That commit created Bite's main files. To "take the change out", Bite deleted them from its copy. But Bite was also installed on the computer, so when the tests ran in the copy, Python quietly found the installed version instead and everything passed. With the install removed, the tests failed like they should. So a test was guarding it. Bite just couldn't see that. That one wrong flag led to two real bugs, which I filed as public issues: 1. The installed-copy leak above (#28). 2. The "check it yourself" command didn't work for new files. Bite prints a command so you can repeat its check by hand, and for a file the commit created, that command just failed (#27). The nice surprise: someone I'd never talked to, bfs2021, fixed #27 in PR #37. Three kinds of false alarm, gone While we were in there, we found two more kinds of false alarm. All three are fixed in PR #42: - Installed copies. Instead of deleting a new file, Bite now swaps in a file that refuses to load, so Python can't sneak in the installed one. - Refactors. A commit called "Refactor ..." or "Rename ..." isn't supposed to change what the code does, so there's nothing for a test to notice. Bite now treats those as tidy-ups and says so. - Docstring edits. Changing only the comments or descriptions in a Python file isn't a code change, so Bite no longer flags it. Each fix came with a test that fails if the fix is taken out. Bite would have been embarrassed otherwise. Then the wording The second problem was me. I read Bite's own results and couldn't follow them. It said things like "TEST GUARDS NOTHING" and "unguarded", and in the pull request comment it put the commit title first. I kept reading commit titles like "Stop three false flags" as instructions to me, not as the thing being checked. So in PR #43 everything Bite shows (the terminal, the report page, the pull request comment) answers that same one question in plain words, verdict first. For example: - ✅ "The test test_high noticed it was gone. If this change breaks later, that test will warn you." - ⚠️ "No test noticed it was gone. If it breaks later, nothing warns you." Then what to do next. !Bit One honest wrinkle: on Bite's own repo, every ✅ names the same test, test_selftest, because Bite's tests all run through one wrapper. It's true, just not very exciting. On normal projects you see the real test names. Does it find anything real? Yes, sometimes. My favourite so far is in text-to-cad. An AI commit added a test meant to catch code that won't run on Python 3.11. Bite noticed the test still passed with the fix taken out. The reason: the way it checked accepted newer syntax, so it passed the exact line it was written to catch. I sent a test that does catch it, PR #708. It was merged and shipped in v0.8.2, and the maintainer then closed the remaining gaps in #712. How often does this happen? One catch is a story, not a number. So this week I had Claude run Bite's free check on 50 public Python projects where people build with AI coding tools. The tests ran in a sealed box with no internet, and it cost nothing. Here's what came out, plainly: - Only 22 of the 50 could run. The rest had tests already failing, needed huge downloads, or took too long. - In those, Bite could check 63 fix claims. For most, a test noticed when the fix was taken out. Good. - For 18, no test noticed. We went through each one by hand. Most were harmless: fixes for the type checker, README edits, code in another language that the Python tests never touch, and one test that only notices outside the UTC timezone (those last two were Bite's mistakes, and they're on my list). - 6 were real. Each one changed how the code behaves, and no test would notice if it vanished. That's about 1 in 10. 5 of the 6 were made with Claude Code. Two honest caveats. All 6 came from just two projects, so it's lumpy, not "every repo has this". And a missing test doesn't mean the fix is broken. It means nothing would warn you if it broke later. For one of those projects, pamqp, I sent a pull request that adds the missing tests, and it was merged (#86). On most projects, Bite stays quiet. Quiet isn't a pass, but it's a good sign. What I took from it - Run your tool on your own work first. Bite's own flag was wrong, and that one wrong flag found two real bugs and three kinds of false alarm. - A wrong alarm costs trust fast. One bad flag and you stop reading the rest. Cutting false alarms mattered more than adding features. - If you can't read your own tool's output, nobody can. I'm the target user who isn't a coder, so when I got confused, that was a bug too. Bite is free to try: pipx install git+https://github.com/connabyss/bite, then bite --free inside a repo. There's also a GitHub Action that leaves that plain-words comment on every pull request. How do you check that an AI's "fixed" is actually fixed? --- Disclosure: I'm not a coder. Claude wrote Bite's code and helped me draft this post. I decide what to build, read every result, and ask until I understand it. The numbers and links above are from the public issues and PRs.

十月第二周 Github 开源热榜：逆向任意程序、PS5 程序移植、AI 端到端测试 #ai工具 #github 1｜REA：给编程 Agent 用的逆向工程 MCP。看到别的软件里想要的功能，让助手不看源码去查：讲清怎么实现、给出证据，再做一个你的版本；能分析原生二进制、JavaScript/Electron、.NET 和网站，分析在本地跑。社区已有人用它做出 ReaCut（重写剪映的开源剪辑器）、Ledger Reborn（本地记账软件）和 zee-rp（让 AI 读写 Axure 原型）；也有实测认为，做简单应用时它比不用 REA 的 Agent 强不了多少。链接：morluto/rea 2｜AnyPS5：把 PS5 可执行文件自动移植成 Linux 和 Windows 程序。靠重链接转成目标系统原生格式，再补上系统库实现，不做模拟；官方网页转换器在浏览器本地完成，文件不离开电脑。链接：boykopovar/AnyPS5 3｜e2e：开源 AI 端到端测试框架，网页和移动端都能测。一句话描述目标让 Agent 操作应用，自然语言步骤和定位、断言写在同一个测试里；验证过的步骤录下来，下次直接重放，不再调用模型。链接：tester-army/e2e 4｜text-to-cad：给 Agent 加上 CAD 能力的插件。本地生成 STEP、GLB、STL、3MF 模型，找现成零件、出工程图，再做钣金和数控评审、切片出 G-code，还能把打印任务发给拓竹打印机。链接：earthtojake/text-to-cad 5｜RAD Debugger：Epic Games 开源的原生图形化调试器，支持 Windows x64 和 PDB，还在 Alpha；同仓库的 RAD Linker 在调试信息几个 G 的大项目上链接快一半，最新版初步支持原生 Linux x64。链接：EpicGames/raddebugger 十月第二周 Github 开源热榜 Top 5，覆盖逆向分析、游戏移植、AI 端到端测试、文字生成 CAD 和原生调试器，Oraink 替你捞重点。 章节： 00:00 开场 · 本周 Top5 00:08 1｜REA 00:59 2｜AnyPS5 01:28 3｜e2e 01:57 4｜text-to-cad 02:26 5｜RAD Debugger 02:58 结尾 · 关注 Oraink

Diseñar PIEZAS 3D con IA es super fácil gracias a text-to-cad Describes la pieza con texto y te genera el archivo listo para imprimir ✓ Funciona con Claude Code / Codex / Cursor ✓ Exporta en STEP / STL / 3MF ✗ Sin saber CAD ➜ earthtojake/text-to-cad 【 100% GRATIS Y OPEN SOURCE 】

text-to-cad Give your agent CAD superpowers.

GitHub’s top: Copy any app, create gorgeous videos, 3d print anything & more Resources: thenextnewthing.ai Presented by Zapier: zapier.com Andrew Warner and Adam explore AI tools that copy apps, reverse-engineer software, create videos and diagrams, and build custom apps and 3D designs. Andrew Warner and Adam break down this week’s top GitHub repos, including REA for reverse engineering apps, Matt Pocock’s developer skills, Hyperframes for AI-generated videos, and Ghidra, the NSA’s decompiler. They also explore tools for creating branded diagrams, extending AI coding sessions with compressed context, generating 3D-printable designs with Text to CAD, improving SwiftUI development, and tracking workouts with an open-source gym app. The big takeaway: AI agents are making powerful developer tools more accessible, while specialized tools still offer greater control, better workflows, and more customization. Repos featured: 1. rea: morluto/rea 2. AnyPS5: boykopovar/AnyPS5 3. Skills by Matt Pocock: mattpocock/skills 4. HyperFrames: heygen-com/hyperframes 5. Diagram Design: cathrynlavery/diagram-design 6. Ghidra: NationalSecurityAgency/ghidra 7. Billion Context: ranxianglei/billion-context 8. Text to CAD: earthtojake/text-to-cad 9. SwiftUI Agent Skill: twostraws/SwiftUI-Agent-Skill 10. openGym: DuarteSantos8/openGym Timestamps: 00:00 - Top GitHub Repos: Apps, AI Videos & Diagrams 00:21 - REA: Copy Apps with AI Agents 03:27 - REA Demo: Rebuilding the 2048 Game 05:11 - PS5 Game Files: Reverse Engineering Without an Emulator 06:29 - Matt Pocock's Skills: Better AI Coding with TDD 09:32 - Why Developers Use Battle-Tested Skills 09:42 - Zapier: Supporting AI Builders 10:10 - Hyperframes: Turn Webpages into Videos 11:41 - Hyperframes Demo: Generate a YouTube Short 12:46 - Branded Diagrams: Create Clean, Custom Visuals 14:40 - Ghidra: The NSA's Decompiler Meets AI Agents 16:07 - REA + Ghidra: Making Reverse Engineering Easier 17:00 - Billion Context: Extend AI Coding Sessions 18:15 - Why Compressed AI Context Can Lose Details 20:31 - Text to CAD: Generate 3D-Printable Designs 21:46 - AI + 3D Printing: Build and Refine Products 23:09 - SwiftUI Agent Skill: Give AI Better Documentation 24:18 - Gym Tracker You Own: An Open-Source Fitness App 25:04 - Why Open-Source Apps Can Beat Ad-Driven Alternatives 26:04 - The Goal: 100,000 Subscribers 👉 Media/Sponsorship Inquiries: thenextnewthing.ai 👉 Join us: thenextnewthing.ai [de-DE] GitHub-Highlights: Beliebige Apps kopieren, beeindruckende Videos erstellen, alles in 3D drucken ... Ressourcen: thenextnewthing.ai Präsentiert von Zapier: zapier.com Andrew Warner und Adam erkunden KI-Tools zum Kopieren von Apps, Reverse Engineering von Software, Erstellen von Videos und Diagrammen sowie zum Entwickeln von benutzerdefinierten Apps und 3D-Designs. Andrew Warner und Adam analysieren die Top-GitHub-Repos dieser Woche, darunter REA für Reverse Engineering von Apps, Matt Pococks Entwicklerfähigkeiten, Hyperframes für KI-generierte Videos und Ghidra, den Dekompiler der NSA. Sie stellen außerdem Tools zur Erstellung von gebrandeten Diagrammen, zur Erweiterung von KI-Codierungssitzungen mit komprimiertem Kontext, zur Generierung von 3D-druckbaren Designs mit Text-to-CAD, zur Verbesserung der SwiftUI-Entwicklung und zum Tracking von Workouts mit einer Open-Source-Fitness-App vor. Fazit: KI-Agenten machen leistungsstarke Entwicklertools zugänglicher, während spezialisierte Tools weiterhin mehr Kontrolle, bessere Workflows und mehr Anpassungsmöglichkeiten bieten. Empfohlene Repositories: 1. rea: morluto/rea 2. AnyPS5: boykopovar/AnyPS5 3. Skills by Matt Pocock: mattpocock/skills 4. HyperFrames: heygen-com/hyperframes 5. Diagram Design: cathrynlavery/diagram-design 6. Ghidra: NationalSecurityAgency/ghidra 7. Billion Context: ranxianglei/billion-context 8. Text to CAD: earthtojake/text-to-cad 9. SwiftUI Agent Skill: twostraws/SwiftUI-Agent-Skill 10. openGym: DuarteSantos8/openGym Zeitstempel: 00:00 – Top-GitHub-Repos: Apps, KI-Videos & Diagramme 00:21 – REA: Apps mit KI-Agenten kopieren 03:27 – REA-Demo: Das Spiel 2048 neu erstellen 05:11 – PS5-Spieldateien: Reverse Engineering ohne Emulator 06:29 – Matt Pococks Fähigkeiten: Bessere KI-Programmierung mit TDD 09:32 – Warum Entwickler auf bewährte Fähigkeiten setzen 09:42 – Zapier: Unterstützung für KI-Entwickler 10:10 – Hyperframes: Webseiten in Videos verwandeln 11:41 – Hyperframes-Demo: Einen YouTube-Kurzfilm erstellen 12:46 – Branded Diagrams: Klare, individuelle Visualisierungen erstellen 14:40 – Ghidra: Der Dekompiler der NSA trifft auf KI-Agenten 16:07 – REA + Ghidra: Reverse Engineering vereinfachen 17:00 – Milliarden Kontext: KI-Codierungssitzungen verlängern 18:15 – Warum komprimierter KI-Kontext Details verlieren kann 20:31 – Text zu CAD: 3D-druckbare Designs generieren 21:46 – KI + 3D-Druck: Produkte entwickeln und optimieren 23:09 – SwiftUI Agent Skill: KI besser dokumentieren 24:18 – Dein eigener Fitness-Tracker: Eine Open-Source-Fitness-App 25:04 – Warum Open-Source-Apps werbefinanzierte Alternativen übertreffen können 26:04 – Das Ziel: 100.000 Abonnenten 👉 Medien-/Sponsoringanfragen: thenextnewthing.ai 👉 Werde Teil unseres Teams: thenextnewthing.ai [es-US] Lo mejor de GitHub: Copia cualquier app, crea videos increíbles, imprime cualquier cosa en 3D y más Recursos: thenextnewthing.ai Presentado por Zapier: zapier.com Andrew Warner y Adam exploran herramientas de IA para copiar aplicaciones, realizar ingeniería inversa de software, crear vídeos y diagramas, y desarrollar aplicaciones personalizadas y diseños 3D. Andrew Warner y Adam analizan los principales repositorios de GitHub de esta semana, incluyendo REA para ingeniería inversa de aplicaciones, las habilidades de desarrollo de Matt Pocock, Hyperframes para vídeos generados por IA y Ghidra, el descompilador de la NSA. También exploran herramientas para crear diagramas personalizados, extender sesiones de codificación de IA con contexto comprimido, generar diseños imprimibles en 3D con Text to CAD, mejorar el desarrollo de SwiftUI y registrar entrenamientos con una aplicación de gimnasio de código abierto. La conclusión principal: los agentes de IA están haciendo que las potentes herramientas de desarrollo sean más accesibles, mientras que las herramientas especializadas siguen ofreciendo mayor control, mejores flujos de trabajo y más opciones de personalización. Repositorios destacados: 1. rea: morluto/rea 2. AnyPS5: boykopovar/AnyPS5 3. Habilidades de Matt Pocock: mattpocock/skills 4. HyperFrames: heygen-com/hyperframes 5. Diseño de diagramas: cathrynlavery/diagram-design 6. Ghidra: NationalSecurityAgency/ghidra 7. Billion Context: ranxianglei/billion-context 8. Texto a CAD: earthtojake/text-to-cad 9. Habilidad de agente de SwiftUI: twostraws/SwiftUI-Agent-Skill 10. openGym: DuarteSantos8/openGym Marcas de tiempo: 00:00 - Principales repositorios de GitHub: Aplicaciones, vídeos y diagramas de IA 00:21 - REA: Copiar aplicaciones con agentes de IA 03:27 - Demo de REA: Reconstruyendo el juego 2048 05:11 - Archivos de juegos de PS5: Ingeniería inversa sin emulador 06:29 - Habilidades de Matt Pocock: Mejor programación de IA con TDD 09:32 - Por qué los desarrolladores usan habilidades probadas 09:42 - Zapier: Apoyando a los creadores de IA 10:10 - Hyperframes: Convierte páginas web en vídeos 11:41 - Demo de Hyperframes: Genera un vídeo corto de YouTube 12:46 - Diagramas personalizados: Crea elementos visuales limpios y personalizados 14:40 - Ghidra: El descompilador de la NSA se une a los agentes de IA 16:07 - REA + Ghidra: Facilitando la ingeniería inversa 17:00 - Contexto de mil millones: Ampliando las sesiones de codificación de IA 18:15 - Por qué el contexto de IA comprimido puede perder detalles 20:31 - Texto a CAD: Generando diseños para impresión 3D 21:46 - IA + Impresión 3D: Creando y perfeccionando productos 23:09 - Habilidad de agente SwiftUI: Mejorando la documentación de la IA 24:18 - Tu propio monitor de actividad física: Una aplicación de fitness de código abierto 25:04 - Por qué las aplicaciones de código abierto pueden superar a las alternativas con publicidad 26:04 - El objetivo: 100.000 suscriptores 👉 Consultas de prensa/patrocinio: thenextnewthing.ai 👉 Únete: thenextnewthing.ai [fr-FR] Le top de GitHub : Copiez n'importe quelle application, créez des vidéos sublimes, imprimez tout ... Ressources : thenextnewthing.ai Présenté par Zapier : zapier.com Andrew Warner et Adam explorent des outils d’IA permettant de copier des applications, de procéder à une ingénierie inverse de logiciels, de créer des vidéos et des diagrammes, et de concevoir des applications personnalisées et des modèles 3D. Andrew Warner et Adam analysent les principaux dépôts GitHub de la semaine, notamment REA pour l’ingénierie inverse d’applications, les compétences de développeur de Matt Pocock, Hyperframes pour les vidéos générées par l’IA et Ghidra, le décompilateur de la NSA. Ils explorent également des outils pour créer des diagrammes personnalisés, enrichir les sessions de programmation IA avec un contexte compressé, générer des modèles imprimables en 3D avec Text to CAD, améliorer le développement SwiftUI et suivre ses entraînements avec une application de fitness open source. Conclusion : les agents d’IA rendent les outils de développement puissants plus accessibles, tandis que les outils spécialisés offrent toujours un meilleur contrôle, des flux de travail optimisés et une personnalisation accrue. Dépôts présentés : 1. rea : morluto/rea 2. AnyPS5 : boykopovar/AnyPS5 3. Skills by Matt Pocock : mattpocock/skills 4. HyperFrames : heygen-com/hyperframes 5. Diagram Design : cathrynlavery/diagram-design 6. Ghidra : NationalSecurityAgency/ghidra 7. Billion Context : ranxianglei/billion-context 8. Text to CAD : earthtojake/text-to-cad 9. SwiftUI Agent Skill : twostraws/SwiftUI-Agent-Skill 10. openGym DuarteSantos8/openGym Horodatage : 00:00 - Meilleurs dépôts GitHub : Applications, vidéos et diagrammes d'IA 00:21 - REA : Copier des applications avec des agents d'IA 03:27 - Démo REA : Recréer le jeu 2048 05:11 - Fichiers de jeu PS5 : Rétro-ingénierie sans émulateur 06:29 - Les compétences de Matt Pocock : Améliorer la programmation d'IA grâce au TDD 09:32 - Pourquoi les développeurs utilisent des compétences éprouvées au combat 09:42 - Zapier : Accompagner les créateurs d'IA 10:10 - Hyperframes : Transformer des pages web en vidéos 11:41 - Démo Hyperframes : Générer une vidéo courte pour YouTube 12:46 - Diagrammes personnalisés : Créer des visuels clairs et personnalisés 14:40 - Ghidra : Quand le décompilateur de la NSA rencontre l'IA Agents 16:07 - REA + Ghidra : Simplifier la rétro-ingénierie 17:00 - Contexte d'un milliard : Prolonger les sessions de programmation IA 18:15 - Pourquoi le contexte IA compressé peut entraîner une perte de détails 20:31 - Texte vers CAO : Générer des conceptions imprimables en 3D 21:46 - IA + Impression 3D : Concevoir et perfectionner des produits 23:09 - Compétence d'agent SwiftUI : Améliorer la documentation de l'IA 24:18 - Suivi d'activité physique : Une application de fitness open source 25:04 - Pourquoi les applications open source peuvent surpasser les alternatives financées par la publicité 26:04 - Objectif : 100 000 abonnés 👉 Demandes médias/parrainage : thenextnewthing.ai 👉 Rejoignez-nous : thenextnewthing.ai [hi] GitHub के टॉप: किसी भी ऐप को कॉपी करें, शानदार वीडियो बनाएं, कुछ भी 3D प्रिंट करें और बहुत कुछ संसाधन: thenextnewthing.ai प्रस्तुतिकर्ता: Zapier: zapier.com एंड्रयू वार्नर और एडम उन AI टूल्स का विश्लेषण करते हैं जो ऐप्स की कॉपी बनाते हैं, सॉफ्टवेयर की रिवर्स इंजीनियरिंग करते हैं, वीडियो और डायग्राम बनाते हैं, और कस्टम ऐप्स और 3D डिज़ाइन तैयार करते हैं। एंड्रयू वार्नर और एडम इस सप्ताह के शीर्ष GitHub रिपॉजिटरी का विस्तृत विवरण देते हैं, जिनमें ऐप्स की रिवर्स इंजीनियरिंग के लिए REA, मैट पोकॉक के डेवलपर कौशल, AI द्वारा निर्मित वीडियो के लिए हाइपरफ्रेम्स और NSA के डीकंपाइलर Ghidra शामिल हैं। वे ब्रांडेड डायग्राम बनाने, संपीड़ित संदर्भ के साथ AI कोडिंग सत्रों को विस्तारित करने, टेक्स्ट टू CAD के साथ 3D प्रिंट करने योग्य डिज़ाइन तैयार करने, SwiftUI डेवलपमेंट को बेहतर बनाने और ओपन-सोर्स जिम ऐप के साथ वर्कआउट को ट्रैक करने के टूल्स का भी विश्लेषण करते हैं। मुख्य निष्कर्ष: AI एजेंट शक्तिशाली डेवलपर टूल्स को अधिक सुलभ बना रहे हैं, जबकि विशेष टूल्स अभी भी बेहतर नियंत्रण, बेहतर वर्कफ़्लो और अधिक अनुकूलन प्रदान करते हैं। प्रदर्शित रिपॉजिटरी: 1. rea: morluto/rea 2. AnyPS5: boykopovar/AnyPS5 3. Skills by Matt Pocock: mattpocock/skills 4. HyperFrames: heygen-com/hyperframes 5. Diagram Design: cathrynlavery/diagram-design 6. Ghidra: NationalSecurityAgency/ghidra 7. Billion Context: ranxianglei/billion-context 8. Text to CAD: earthtojake/text-to-cad 9. SwiftUI Agent Skill: twostraws/SwiftUI-Agent-Skill 10. openGym: DuarteSantos8/openGym टाइमस्टैम्प: 00:00 - टॉप GitHub रिपॉज़िटरी: ऐप्स, AI वीडियो और डायग्राम 00:21 - REA: AI एजेंट के साथ ऐप्स कॉपी करें 03:27 - REA डेमो: 2048 गेम को रीबिल्ड करना 05:11 - PS5 गेम फ़ाइलें: एमुलेटर के बिना रिवर्स इंजीनियरिंग 06:29 - मैट पोकॉक के कौशल: TDD के साथ बेहतर AI कोडिंग 09:32 - डेवलपर्स क्यों आजमाए हुए कौशल का उपयोग करते हैं 09:42 - Zapier: AI बिल्डर्स को सपोर्ट करना 10:10 - हाइपरफ्रेम: वेबपेज को वीडियो में बदलें 11:41 - हाइपरफ्रेम डेमो: एक YouTube शॉर्ट बनाएं 12:46 - ब्रांडेड डायग्राम: साफ़, कस्टम विज़ुअल बनाएं 14:40 - Ghidra: NSA का डीकंपाइलर AI एजेंट से मिलता है 16:07 - REA + Ghidra: रिवर्स इंजीनियरिंग को आसान बनाना 17:00 - बिलियन कॉन्टेक्स्ट: AI कोडिंग सेशन को विस्तारित करना 18:15 - संपीड़ित AI कॉन्टेक्स्ट में विवरण क्यों खो सकते हैं 20:31 - टेक्स्ट से CAD: 3D प्रिंट करने योग्य डिज़ाइन तैयार करना 21:46 - AI + 3D प्रिंटिंग: उत्पाद बनाना और परिष्कृत करना 23:09 - SwiftUI एजेंट स्किल: AI को बेहतर दस्तावेज़ीकरण प्रदान करना 24:18 - आपका अपना जिम ट्रैकर: एक ओपन-सोर्स फिटनेस ऐप 25:04 - ओपन-सोर्स ऐप्स विज्ञापन-आधारित विकल्पों को कैसे मात दे सकते हैं 26:04 - लक्ष्य: 100,000 सब्सक्राइबर 👉 मीडिया/प्रायोजन संबंधी पूछताछ: thenextnewthing.ai 👉 हमसे जुड़ें: thenextnewthing.ai [id] Top GitHub: Salin aplikasi apa pun, buat video yang memukau, cetak 3D apa pun & lainnya Sumber daya: thenextnewthing.ai Dipersembahkan oleh Zapier: zapier.com Andrew Warner dan Adam mengeksplorasi alat AI yang menyalin aplikasi, merekayasa balik perangkat lunak, membuat video dan diagram, serta membangun aplikasi khusus dan desain 3D. Andrew Warner dan Adam membahas repositori GitHub teratas minggu ini, termasuk REA untuk merekayasa balik aplikasi, keterampilan pengembang Matt Pocock, Hyperframes untuk video yang dihasilkan AI, dan Ghidra, dekompiler NSA. Mereka juga mengeksplorasi alat untuk membuat diagram bermerek, memperluas sesi pengkodean AI dengan konteks terkompresi, menghasilkan desain yang dapat dicetak 3D dengan Text to CAD, meningkatkan pengembangan SwiftUI, dan melacak latihan dengan aplikasi gym sumber terbuka. Kesimpulan utamanya: agen AI membuat alat pengembang yang canggih lebih mudah diakses, sementara alat khusus masih menawarkan kontrol yang lebih besar, alur kerja yang lebih baik, dan lebih banyak penyesuaian. Repositori yang ditampilkan: 1. rea: morluto/rea 2. AnyPS5: boykopovar/AnyPS5 3. Skills by Matt Pocock: mattpocock/skills 4. HyperFrames: heygen-com/hyperframes 5. Diagram Design: cathrynlavery/diagram-design 6. Ghidra: NationalSecurityAgency/ghidra 7. Billion Context: ranxianglei/billion-context 8. Text to CAD: earthtojake/text-to-cad 9. SwiftUI Agent Skill: twostraws/SwiftUI-Agent-Skill 10. openGym: DuarteSantos8/openGym Cap waktu: 00:00 - Repositori GitHub Teratas: Aplikasi, Video AI & Diagram 00:21 - REA: Menyalin Aplikasi dengan Agen AI 03:27 - Demo REA: Membangun Kembali Game 2048 05:11 - File Game PS5: Rekayasa Balik Tanpa Emulator 06:29 - Keterampilan Matt Pocock: Pengkodean AI yang Lebih Baik dengan TDD 09:32 - Mengapa Pengembang Menggunakan Keterampilan yang Teruji di Medan Perang 09:42 - Zapier: Mendukung Pembuat AI 10:10 - Hyperframes: Mengubah Halaman Web menjadi Video 11:41 - Demo Hyperframes: Menghasilkan Video Pendek YouTube 12:46 - Branded Diagrams: Membuat Visual Kustom yang Bersih 14:40 - Ghidra: Dekompiler NSA Bertemu Agen AI 16:07 - REA + Ghidra: Melakukan Rekayasa Balik Lebih Mudah 17:00 - Konteks Miliaran: Memperpanjang Sesi Pengkodean AI 18:15 - Mengapa Konteks AI Terkompresi Dapat Kehilangan Detail 20:31 - Teks ke CAD: Menghasilkan Desain yang Dapat Dicetak 3D 21:46 - AI + Pencetakan 3D: Membangun dan Memperbaiki Produk 23:09 - Keterampilan Agen SwiftUI: Memberikan Dokumentasi yang Lebih Baik untuk AI 24:18 - Pelacak Gym yang Anda Miliki: Aplikasi Kebugaran Sumber Terbuka 25:04 - Mengapa Aplikasi Sumber Terbuka Dapat Mengalahkan Alternatif Berbasis Iklan 26:04 - Tujuan: 100.000 Pelanggan 👉 Pertanyaan Media/Sponsor: thenextnewthing.ai 👉 Bergabunglah dengan kami: thenextnewthing.ai [it] I migliori progetti di GitHub: copia qualsiasi app, crea video spettacolari, stampa in 3D qualsia... Risorse: thenextnewthing.ai Presentato da Zapier: zapier.com Andrew Warner e Adam esplorano strumenti di intelligenza artificiale per copiare app, eseguire il reverse engineering di software, creare video e diagrammi e realizzare app personalizzate e progetti 3D. Andrew Warner e Adam analizzano i migliori repository GitHub di questa settimana, tra cui REA per il reverse engineering di app, le competenze di sviluppo di Matt Pocock, Hyperframes per video generati dall'IA e Ghidra, il decompilatore della NSA. Esplorano anche strumenti per creare diagrammi personalizzati, estendere le sessioni di programmazione IA con il contesto compresso, generare progetti stampabili in 3D con Text to CAD, migliorare lo sviluppo con SwiftUI e monitorare gli allenamenti con un'app open source per la palestra. Il messaggio principale: gli agenti IA stanno rendendo più accessibili potenti strumenti di sviluppo, mentre gli strumenti specializzati offrono ancora maggiore controllo, flussi di lavoro migliori e maggiori possibilità di personalizzazione. Repository in evidenza: 1. rea: morluto/rea 2. AnyPS5: boykopovar/AnyPS5 3. Skills by Matt Pocock: mattpocock/skills 4. HyperFrames: heygen-com/hyperframes 5. Diagram Design: cathrynlavery/diagram-design 6. Ghidra: NationalSecurityAgency/ghidra 7. Billion Context: ranxianglei/billion-context 8. Text to CAD: earthtojake/text-to-cad 9. SwiftUI Agent Skill: twostraws/SwiftUI-Agent-Skill 10. openGym: Italiano: DuarteSantos8/openGym Timestamp: 00:00 - I migliori repository GitHub: app, video e diagrammi sull'IA 00:21 - REA: copiare app con agenti IA 03:27 - Demo di REA: ricostruire il gioco 2048 05:11 - File di gioco PS5: reverse engineering senza emulatore 06:29 - Le competenze di Matt Pocock: migliorare la programmazione IA con TDD 09:32 - Perché gli sviluppatori utilizzano competenze collaudate 09:42 - Zapier: supporto per i creatori di IA 10:10 - Hyperframes: trasformare le pagine web in video 11:41 - Demo di Hyperframes: generare un breve video per YouTube 12:46 - Diagrammi personalizzati: creare elementi visivi puliti e personalizzati 14:40 - Ghidra: il decompiler della NSA incontra gli agenti IA 16:07 - REA + Ghidra: Semplificare il reverse engineering 17:00 - Billion Context: Estendere le sessioni di programmazione AI 18:15 - Perché il contesto AI compresso può perdere dettagli 20:31 - Da testo a CAD: Generare progetti stampabili in 3D 21:46 - AI + Stampa 3D: Creare e perfezionare prodotti 23:09 - SwiftUI Agent Skill: Fornire all'AI una documentazione migliore 24:18 - Gym Tracker You Own: Un'app fitness open source 25:04 - Perché le app open source possono battere le alternative basate sulla pubblicità 26:04 - L'obiettivo: 100.000 iscritti 👉 Richieste media/sponsorizzazioni: thenextnewthing.ai 👉 Unisciti a noi: thenextnewthing.ai [pt-BR] Os melhores do GitHub: copie qualquer app, crie vídeos incríveis, imprima qualquer coisa em 3D e ... Recursos: thenextnewthing.ai Apresentado por Zapier: zapier.com Andrew Warner e Adam exploram ferramentas de IA que copiam aplicativos, fazem engenharia reversa de software, criam vídeos e diagramas e desenvolvem aplicativos personalizados e designs 3D. Andrew Warner e Adam analisam os principais repositórios do GitHub desta semana, incluindo o REA para engenharia reversa de aplicativos, as habilidades de desenvolvimento de Matt Pocock, o Hyperframes para vídeos gerados por IA e o Ghidra, o descompilador da NSA. Eles também exploram ferramentas para criar diagramas personalizados, estender sessões de codificação de IA com contexto compactado, gerar designs imprimíveis em 3D com o Text to CAD, aprimorar o desenvolvimento em SwiftUI e monitorar exercícios com um aplicativo de academia de código aberto. A principal conclusão: os agentes de IA estão tornando as poderosas ferramentas de desenvolvimento mais acessíveis, enquanto as ferramentas especializadas ainda oferecem maior controle, melhores fluxos de trabalho e mais opções de personalização. Andrew Warner e Adam analisam os principais repositórios do GitHub desta semana, incluindo o REA para engenharia reversa de aplicativos, as ferramentas especializadas para desenvolvimento de software, o Hyperframes para vídeos gerados por IA e o Ghidra, o descompilador da NSA. Eles também exploram ferramentas para criar diagramas personalizados, estender sessões de codificação de IA com contexto compactado, gerar designs imprimíveis em 3D com o Text to CAD, aprimorar o desenvolvimento em SwiftUI e monitorar exercícios com um aplicativo de academia de código aberto. A principal conclusão: os agentes de IA estão tornando as ferramentas de desenvolvimento mais acessíveis, enquanto as ferramentas especializadas ainda oferecem maior controle, melhores fluxos de trabalho e mais opções de personalização. Andrew Warner e Adam analisam os principais repositórios do GitHub desta semana. Repositórios em destaque: 1. rea: morluto/rea 2. AnyPS5: boykopovar/AnyPS5 3. Skills by Matt Pocock: mattpocock/skills 4. HyperFrames: heygen-com/hyperframes 5. Diagram Design: cathrynlavery/diagram-design 6. Ghidra: NationalSecurityAgency/ghidra 7. Billion Context: ranxianglei/billion-context 8. Text to CAD: earthtojake/text-to-cad 9. SwiftUI Agent Skill: twostraws/SwiftUI-Agent-Skill 10. openGym: DuarteSantos8/openGym Marcadores de tempo: 00:00 - Principais repositórios do GitHub: Aplicativos, vídeos de IA e diagramas 00:21 - REA: Copie aplicativos com agentes de IA 03:27 - Demonstração do REA: Reconstruindo o jogo 2048 05:11 - Arquivos de jogos do PS5: Engenharia reversa sem um emulador 06:29 - Habilidades de Matt Pocock: Melhor programação de IA com TDD 09:32 - Por que os desenvolvedores usam habilidades testadas e aprovadas 09:42 - Zapier: Apoiando criadores de IA 10:10 - Hyperframes: Transforme páginas da web em vídeos 11:41 - Demonstração do Hyperframes: Gere um curta para o YouTube 12:46 - Branded Diagrams: Crie visuais limpos e personalizados 14:40 - Ghidra: O descompilador da NSA encontra agentes de IA 16:07 - REA + Ghidra: Facilitando a Engenharia Reversa 17:00 - Billion Context: Estenda as Sessões de Programação de IA 18:15 - Por que o Contexto de IA Comprimido Pode Perder Detalhes 20:31 - Texto para CAD: Gere Projetos Imprimíveis em 3D 21:46 - IA + Impressão 3D: Construa e Aprimore Produtos 23:09 - Habilidade de Agente SwiftUI: Dê à IA uma Documentação Melhor 24:18 - Gym Tracker Seu: Um Aplicativo de Fitness de Código Aberto 25:04 - Por que Aplicativos de Código Aberto Podem Superar Alternativas Baseadas em Anúncios 26:04 - A Meta: 100.000 Inscritos 👉 Consultas de Mídia/Patrocínio: thenextnewthing.ai 👉 Junte-se a nós: thenextnewthing.ai

Text-to-CAD: Open-Source AI CAD Generator for Claude Code Text-to-CAD is an MIT-licensed plugin that gives Claude Code, Codex, Cursor, Gemini and Grok a full CAD workflow. Instead of generating a mesh, your agent writes a parametric Python model that produces exact STEP geometry, then exports STL or 3MF, draws dimensioned drawings, checks manufacturability and sends parts to a printer. What Text-to-CAD Is: CAD Superpowers for Your AI Coding Agent Text-to-CAD is an open-source plugin with a one-line promise: "Give your agent CAD superpowers." Install it in Claude Code, Codex, Cursor, Gemini or Grok Build, describe a part in plain language or show it a sketch, and your agent produces a real 3D model. It goes well past a single model, too. The plugin gives the agent local workflows to: - generate 3D models as STEP, STL, 3MF or GLB files - produce dimensioned engineering drawings as PDF - run design-for-manufacturing checks for sheet metal, CNC machining, injection molding and 3D printing - find off-the-shelf parts like screws, bearings and motors as ready-made STEP files - slice models into G-code and send them to 3D printers, or check files before ordering laser-cut parts - write robot description files for ROS and simulators It is built on two serious open-source foundations: build123d 0.11, a Python CAD library, and Open CASCADE 7.9, the geometry kernel behind FreeCAD. The engine ships on PyPI as cadgen. The project is MIT licensed, had 18,586 stars at the time of writing, and released v0.7.19 on 9 October 2026. One naming note: this is not Zoo's Text-to-CAD, a separate commercial product with the same name. This one is a free plugin, created by earthtojake, that runs locally inside the coding agent you already use. CAD as Code: Why It Writes Python Instead of Generating a Mesh This is the idea that separates Text-to-CAD from most "AI 3D model generators". Typical text-to-3D tools output a mesh: thousands of triangles approximating a surface. That is fine for a figurine. It is useless for engineering, because a mesh has no true circles, no exact hole diameters and no parameters. You cannot reliably change a 6 mm wall to 8 mm in a triangle soup. Text-to-CAD has the agent write code instead. A model is a plain Python script with a decorated function that returns a build123d shape: from cadgen import build123d as bd from cadgen import step WIDTH = 40.0 @step(out="../STEP/bracket.step") def bracket(): body = bd.Box(WIDTH, 20, 6) body.label = "bracket" return body Running the script writes a STEP file: exact boundary-representation geometry, the universal exchange format that FreeCAD, Fusion and SolidWorks all import. Want the bracket wider? Change WIDTH and rerun. Want a 3D-printable copy? Stack an @stl or @threemf decorator on the same function, and the mesh is derived from the exact model rather than the other way round. Because the model is source code, everything good about code applies: it lives in git, a change shows up as a readable diff, and assemblies are built by calling child models. The skill enforces reproducibility too. Geometry "must not depend on untracked time, random values, environment variables or the working directory", so the same script always builds the same part. The Agent Has to Prove the Part Is Right An AI that says "done" is not the same as a part that fits. The CAD skill turns verification into an explicit step of the workflow. After building, the agent is told to choose checks from the requested dimensions, clearances and topology, and to run them against the saved file, not the source. Measurements use native build123d geometry. Then it must render at least one snapshot of the result and read it itself, with the instruction to "always render and read them yourself, never rely on the viewer for it." Two lines from the skill set the tone for the whole project: > A failed computation is not a pass. > Report units, thresholds, selected geometry and untested requirements. The hand-off has to list output files, the checks actually run, and any assumptions. For unfamiliar dimensions, the agent records what it assumed. When missing information would materially change the result, it asks you instead of guessing. Failures are repaired in the source, and the affected checks rerun. If you are working from a spec sheet in inches while the models default to millimeters, a unit converter settles the numbers before they reach the prompt. Point at a Face in the Viewer and Ask for a Change The plugin includes CAD Viewer, a local server that shows models where you are working. In Claude Desktop, models appear as viewer cards in the chat that you can orbit and open full size. In the Codex app there is a CAD tab beside each thread that the agent drives. In terminal agents, asking to see a model gives you a link that opens it in your browser. The clever part is that selection flows back into the conversation. Click a face and the viewer copies a precise reference to it, such as assembly.step#o1.2.f7, which pins down that one face in that saved file. The agent opens the document with read_scene, resolves the reference to real geometry, and knows exactly which surface "this face" means. A Quick Edit goes further: you describe the change, the viewer attaches the file and references, and if you sketched on the view, a PNG of your markup. The agent is instructed to look at that sketch before touching the model. It is the CAD equivalent of circling something on a printout and writing "make this thicker" next to it. The agent also has its own eyes: cad_screenshot shows it what you are seeing, and cad_view reads what you have selected. Twelve Skills, From Bolts to Bambu Lab Printers Text-to-CAD is a library of focused skills. Install the plugin and the agent loads whichever one a request needs: | Skill | What it does | |---|---| | CAD | Creates and edits models from text or images; STEP first, plus STL, 3MF and GLB | | step.parts | Finds off-the-shelf STEP parts: screws, bearings, motors, connectors | | Engineering Drawing | Dimensioned PDF drawings with views, hidden lines, hole callouts and a title block | | DXF | 2D profiles, templates, gaskets and cut layouts | | URDF | Robot structure files: links, joints, limits, inertials, meshes | | SRDF | MoveIt planning groups, end effectors, poses and collision rules | | SDF | Simulator models and worlds with physics, sensors and lights | | SendCutSend | Checks DXF and STEP files before upload to SendCutSend | | DfAM Check | 3D-printability: wall thickness, overhangs, support volume, orientation | | DFM | Sheet metal, CNC and injection molding review with cited rules | | G-code | Slices models with OrcaSlicer using your own printer presets | | Bambu Labs | Sends prints to Bambu Lab printers via Bambu Connect or Bambu Studio | The breadth matters because real projects cross these lines. A robot arm needs CAD parts, purchased servos, a URDF for ROS, and printed brackets. Text-to-CAD keeps the whole chain in one conversation. Real Parts and Real Manufacturing Rules, Not Placeholders Two skills keep designs grounded in the physical world. step.parts is a hosted catalog of purchasable components with an API. When an assembly needs a named servo, motor or fastener, the agent is told to search it before inventing a placeholder, to try aliases and vendor names before giving up, and to verify each downloaded STEP file against its SHA-256 hash when the catalog publishes one. If it does fall back to a simplified envelope, it has to say so. DFM reviews a part for sheet metal, CNC machining or injection molding, and its rules are strict about evidence: - It is "a guided review skill, not an automatic feature-recognition or manufacturing certification engine", and unavailable checks must be reported as unavailable. - Your actual supplier's specification beats general guidance, and conflicting specifications are recorded rather than silently resolved. - Every limit adopted must cite its source, version, access date and units. If a rule cannot be checked, the review reports it missing rather than making up a default. - It must never infer an alloy, resin, strength or stock thickness from a rendering's material or color. - For injection molding, a script measures draft, undercuts and projected area from the mesh. 3D printing has its own skill, DfAM Check, which measures printability per process, and the G-code and Bambu Lab skills take a checked model all the way to a print. Robots Too: URDF, SRDF and SDF Files Robotics is where Text-to-CAD quietly saves the most tedium. Robot description files are verbose, unforgiving and usually written by hand. - URDF describes a robot's links and joints, with limits, inertials and meshes, and is the standard format in ROS. - SRDF adds MoveIt's semantic layer: planning groups, end effectors, named poses and which collisions to ignore. - SDF describes models and whole worlds for simulators, including physics, sensors and lights. Because the parts themselves come from the CAD skill, the meshes in these files can come from the real geometry rather than eyeballed stand-ins. All three formats are XML; when you want to inspect one programmatically or feed it into a web tool, an XML to JSON converter gives you a structure that is much easier to query. How to Install Text-to-CAD in Claude Code, Codex or Cursor The simplest install is to send your agent one message: Install text-to-cad from https://github.com/earthtojake/text-to-cad To do it yourself, first make sure uv, Astral's Python package manager, is installed (uv --version). Then, in Claude Code: claude plugin marketplace add earthtojake/text-to-cad#latest claude plugin install text-to-cad@earthtojake Restart the app; the first start downloads the CAD runtime, so it needs a network connection. Codex installs from its plugin directory, Gemini as an extension with --auto-update, and Cursor and Grok Build can reuse the Claude Code plugin directly. Any agent that supports the skills framework can install the skills alone with npx skills add earthtojake/text-to-cad#latest. Three tips from the README: - Install the plugin or the skills, not both, or every skill loads twice. - The commands use the latest branch, which only receives released versions. Leaving it out installs main, which can contain unreleased work. - When updating skills, rerun npx skills add rather than npx skills update, which silently misses newly added skills. Windows 11 gotcha. The CAD kernel's Python binding ships an unsigned native module, and Windows 11's Smart App Control, on by default on fresh installs, blocks it. Every command then fails with ImportError: DLL load failed while importing OCP. There is no per-app exception: either turn Smart App Control off, which can only be reversed by reinstalling Windows, or run the CAD skills under WSL. cadgen doctor detects this case. Telemetry Is On by Default: Here Is Exactly What It Sends Text-to-CAD sends usage statistics and crash reports by default, and the README documents them in unusual detail, so it is worth knowing before you install. What is sent: versions, install source, operating system and agent app, and counts, such as how often each CAD tool was called, how many models were built, and why a call failed, chosen from a fixed set of reasons rather than the error message. Crash reports include the error type and where in cadgen it failed. Everything is tagged with a random install ID and stored with PostHog. The server derives your country from your IP address and does not keep the address. What is never sent: file names, paths, file contents or prompts. How to turn it off: run uvx cadgen telemetry off (which also deletes what was already sent), use the menu in the viewer, or ask your agent. DO_NOT_TRACK=1 disables it for a single process. Nothing is sent in CI or from a development install, and everything works offline. Separately, cadgen checks for a new version at most once a day with one anonymous request; CADGEN_UPDATE_CHECK=0 stops that too. Repo Health: 18,586 Stars and a Fast Release Cadence The repository was created on 22 April 2026 and had 18,586 stars, 1,848 forks, 1,453 commits and 20 contributors at the time of writing. Development is fast: v0.7.17, v0.7.18 and v0.7.19 all shipped between 8 and 9 October 2026, and the cadgen package has had 57 releases on PyPI. A test workflow runs in CI, documentation lives at texttocad.dev, and there is a Discord community. Good fit if you design functional parts: brackets, enclosures, mounts, jigs, robot components or anything with dimensions that matter. It suits makers with a 3D printer, robotics developers tired of hand-writing URDF, and engineers who want a parametric starting point they can open in their usual CAD tool. Poor fit if you want organic or sculpted shapes such as characters and figurines, where mesh-based text-to-3D generators are the better tool. It is also not a substitute for a manufacturer's sign-off: the DFM skill is explicitly a guided review, not certification. And on Windows 11, plan for the Smart App Control step above. If you like the idea of agents doing serious engineering work, the mirror image of this project is REA, which points an agent at existing software and takes it apart instead. FAQ What is Text-to-CAD? Text-to-CAD is a free, MIT-licensed plugin that gives AI coding agents like Claude Code, Codex, Cursor, Gemini and Grok a full CAD workflow. The agent writes parametric Python models with build123d that produce STEP files, and can export STL, 3MF and GLB, generate engineering drawings, run manufacturability checks and send parts to printers. Is Text-to-CAD free? Yes. The plugin and its cadgen engine are open source under the MIT license. You only need an AI agent you already use and uv to run it. Fabrication services and printers it connects to are paid separately as usual. Is this the same as Zoo's Text-to-CAD? No. Zoo's Text-to-CAD is a separate commercial product. earthtojake/text-to-cad is an independent open-source plugin that runs locally inside your coding agent. What file formats does Text-to-CAD produce? STEP is the main output, giving exact geometry you can open in FreeCAD, Fusion or SolidWorks. It also exports STL, 3MF and GLB meshes, DXF for 2D profiles, PDF engineering drawings, G-code for printers, and URDF, SRDF and SDF files for robotics. Which AI agents support Text-to-CAD? It has plugins or extensions for Claude Code, Claude Desktop, Codex, Cursor, Grok Build and Gemini, and any agent that supports the skills framework can install the skills with npx skills add. Can I edit the models in other CAD software? Yes. Models are saved as STEP files, the standard exchange format for mechanical CAD, so FreeCAD, Fusion, SolidWorks and most other CAD tools can import them. The Python source also stays editable, so changing a parameter and rerunning rebuilds the part. Does Text-to-CAD upload my designs? No. Models are built locally. By default it sends anonymous usage statistics and crash reports, but never file names, paths, contents or prompts, and you can turn that off with uvx cadgen telemetry off. Why does Text-to-CAD fail with 'DLL load failed while importing OCP' on Windows? Windows 11's Smart App Control blocks the unsigned native module in the CAD kernel's Python binding. Turn Smart App Control off, which can only be undone by reinstalling Windows, or run the CAD skills under WSL. Running cadgen doctor confirms the cause. --- Originally published on DevPik.

이번 주 개발자가 놓치면 안 될 GitHub 트렌딩 Top 10 총정리 (10월 2주) 이번 주(10/7~10/14) GitHub 트렌딩 주간 순위 Top 10과 순위표 밖 저장소 3개를, 레포마다 무엇인지·이번 주 바뀐 것·주의할 점까지 자세히 정리했어요. 순위는 GitHub Trending(weekly) 페이지 순서 그대로예요(2026-10-14 06:00 KST 수집). 별 수는 수집 시점 값이에요. 1위 AnyPS5 — PS5 실행 파일을 에뮬레이터 없이 Linux·Windows용으로 다시 링크하는 도구 (GPL-2.0). 영상 속 게임 화면(Animal Well·Ape Academy 2·Astro Bot·Balatro·Dreaming Sarah·Hello Neighbor 2)은 저장소 README에 올라온 Windows 실행 화면이에요. ⚠ 게임 실행 파일을 어떻게 얻는지는 문서에 없고, 합법성은 사용자 책임으로 넘깁니다. 복제 방지 우회(DMCA 등) 문제가 지적돼 짧게만 다뤘어요. 2위 mattpocock/skills — 되묻기 → 스펙 → 티켓 → TDD → 리뷰 순서로 일하게 하는 코딩 에이전트 스킬 모음 (MIT) 3위 tester-army/e2e — 자연어 단계를 섞어 쓰는 E2E 테스트 (0.19.0 toHaveScreenshot, 0.20.0 run --watch) 4위 EpicGames/raddebugger — 여러 프로세스를 함께 다루는 네이티브 디버거 (ALPHA, 리눅스 예비 지원) 5위 heygen-com/hyperframes — HTML로 쓰고 영상으로 렌더 (transcribe --to srt, 텔레메트리 끄기 무시 경로 수정 v0.8.151) 6위 thedotmack/claude-mem — 코딩 에이전트 세션 간 기억 (별 10만 돌파, v13.37.0부터 제공자 실패 시 기본으로 사용자의 Anthropic 플랜 사용) 7위 cursor/plugins — Cursor 플러그인 규격·공식 플러그인 (Workday·Salesforce·Square 연동 추가) 8위 earthtojake/text-to-cad — 말로 3D 부품 설계 (10/8부터 텔레메트리 기본 켜짐) 9위 flutter/flutter — 3.47.7 핫픽스: iOS 접근성 버그 2건 수정 10위 cathrynlavery/diagram-design — 말로 시키면 구성도를 HTML+SVG 한 파일로 (45종) 순위 밖 morluto/rea — 리버스 엔지니어링 도구에 에이전트를 연결 (별 8,692 → 203,368, 10/7→10/14 GitHub API · HN 797점 10/10 · 주간 트렌딩 페이지에는 없음) ▶ 텔레메트리(사용 기록 전송) 끄는 법 - text-to-cad: uvx cadgen telemetry off 또는 DO_NOT_TRACK=1 - claude-mem: npx claude-mem telemetry disable · 요금제 자동 전환 막기 CLAUDE_MEM_QUOTA_FALLBACK_PROVIDER=off - hyperframes: HYPERFRAMES_NO_TELEMETRY=1 (v0.8.151 이상으로 업데이트) - e2e: E2E_TELEMETRY_DISABLED=1 - Flutter: flutter --disable-analytics ▶ 이번 주 함께 볼 것 - vercel-labs/skills(npx skills) v1.7.2: package.json의 skills 필드로 팀 공용 스킬 설치·정리(sync). 텔레메트리 기본 켜짐(DISABLE_TELEMETRY=1) - debpalash/VoiceStudio: 원격 기기·워커 기능을 유료 Pro로 이전(10/11), 상업 라이선스가 불분명한 기본 음성 모델은 동의 후 사용(10/10) ▶ 챕터 0:00 오프닝 0:20 10위 diagram-design 0:59 9위 Flutter 1:27 8위 text-to-cad 2:00 7위 cursor/plugins 2:33 6위 claude-mem 3:09 5위 HyperFrames 3:39 4위 RAD Debugger 4:06 3위 e2e 4:52 2위 mattpocock/skills 5:36 순위 밖 ① REA 6:16 순위 밖 ② vercel-labs/skills 6:43 순위 밖 ③ VoiceStudio 7:07 1위 AnyPS5 7:48 이번 주 확인할 설정 ▶ 관련 영상 - 이번 주 트렌딩 쇼츠 요약판: youtu.be - REA 직접 써 본 영상: youtu.be - e2e 실측: youtu.be ▶ 링크 boykopovar/AnyPS5 mattpocock/skills tester-army/e2e EpicGames/raddebugger heygen-com/hyperframes thedotmack/claude-mem cursor/plugins earthtojake/text-to-cad flutter/flutter cathrynlavery/diagram-design morluto/rea ※ 내레이션은 AI 보이스(ElevenLabs)예요. 실화면이 없는 장면과 설명용 그림은 화면에 "예)"로 표시했고, 공식 문서에 적힌 동작 순서를 그린 원리 모션이에요. #GitHub트렌딩 #오픈소스 #개발자 #AI에이전트 #코딩에이전트 #SoloStack [en] GitHub Trending Top 10 This Week, Explained (Oct, Week 2) The weekly GitHub Trending Top 10 (collected 2026-10-14 06:00 KST) plus three repos outside the chart, each with what it is, what changed this week, and what to watch out for (licenses, telemetry defaults). Narration is an AI voice (ElevenLabs). Explanatory illustrations are marked "예)" (example).

이번 주 개발자가 놓치면 안 될 GitHub 트렌딩 Top 10 (10월 2주) 이번 주(10/7~10/14) GitHub 트렌딩 주간 순위 Top 10과, 순위표 밖에서 가장 뜨거웠던 저장소 하나를 정리했어요. 순위는 GitHub Trending(weekly) 페이지 순서 그대로예요(2026-10-14 06:00 KST 수집). 별 수는 수집 시점 값이에요. 1위 AnyPS5 — PS5 실행 파일을 에뮬레이터 없이 Linux·Windows용으로 다시 링크하는 도구 (GPL-2.0). 영상 속 게임 화면(Animal Well·Ape Academy 2·Astro Bot·Balatro·Dreaming Sarah·Hello Neighbor 2)은 저장소 README에 올라온 Windows 실행 화면이에요. ⚠ 게임 실행 파일을 어떻게 얻는지는 문서에 없고, 합법성은 사용자 책임으로 넘깁니다. 복제 방지 우회(DMCA 등) 문제가 지적돼 짧게만 다뤘어요. 2위 mattpocock/skills — 되묻기 → 스펙 → 티켓 → TDD → 리뷰 순서로 일하게 하는 코딩 에이전트 스킬 모음 (MIT) 3위 tester-army/e2e — 자연어 단계를 섞어 쓰는 E2E 테스트 (0.19.0 toHaveScreenshot, 0.20.0 run --watch) 4위 EpicGames/raddebugger — 여러 프로세스를 함께 다루는 네이티브 디버거 (ALPHA, 리눅스 예비 지원) 5위 heygen-com/hyperframes — HTML로 쓰고 영상으로 렌더 (transcribe --to srt, 텔레메트리 끄기 무시 경로 수정 v0.8.151) 6위 thedotmack/claude-mem — 코딩 에이전트 세션 간 기억 (별 10만 돌파, v13.37.0부터 제공자 실패 시 기본으로 사용자의 Anthropic 플랜 사용) 7위 cursor/plugins — Cursor 플러그인 규격·공식 플러그인 (Workday·Salesforce·Square 연동 추가) 8위 earthtojake/text-to-cad — 말로 3D 부품 설계 (10/8부터 텔레메트리 기본 켜짐) 9위 flutter/flutter — 3.47.7 핫픽스: iOS 접근성 버그 2건 수정 10위 cathrynlavery/diagram-design — 말로 시키면 구성도를 HTML+SVG 한 파일로 (45종) 순위 밖 morluto/rea — 리버스 엔지니어링 도구에 에이전트를 연결 (별 8,692 → 203,368, 10/7→10/14 GitHub API · HN 797점 10/10 · 주간 트렌딩 페이지에는 없음) ▶ 텔레메트리(사용 기록 전송) 끄는 법 - text-to-cad: uvx cadgen telemetry off 또는 DO_NOT_TRACK=1 - claude-mem: npx claude-mem telemetry disable · 요금제 자동 전환 막기 CLAUDE_MEM_QUOTA_FALLBACK_PROVIDER=off - hyperframes: HYPERFRAMES_NO_TELEMETRY=1 (v0.8.151 이상으로 업데이트) - e2e: E2E_TELEMETRY_DISABLED=1 - Flutter: flutter --disable-analytics ▶ 이번 주 함께 볼 것 - vercel-labs/skills(npx skills) v1.7.2: package.json의 skills 필드로 팀 공용 스킬 설치·정리(sync). 텔레메트리 기본 켜짐(DISABLE_TELEMETRY=1) - debpalash/VoiceStudio: 원격 기기·워커 기능을 유료 Pro로 이전(10/11), 상업 라이선스가 불분명한 기본 음성 모델은 동의 후 사용(10/10) ▶ 관련 영상 - 이번 주 트렌딩 롱폼 총정리(레포마다 자세히): youtu.be - REA 직접 써 본 영상: youtu.be - e2e 실측: youtu.be ▶ 링크 boykopovar/AnyPS5 mattpocock/skills tester-army/e2e EpicGames/raddebugger heygen-com/hyperframes thedotmack/claude-mem cursor/plugins earthtojake/text-to-cad flutter/flutter cathrynlavery/diagram-design morluto/rea ※ 내레이션은 AI 보이스(ElevenLabs)예요. 실화면이 없는 저장소(text-to-cad·RAD Debugger·mattpocock/skills·AnyPS5)는 공식 문서에 적힌 동작 순서를 그린 원리 모션이에요. #GitHub트렌딩 #오픈소스 #개발자 #AI에이전트 #코딩에이전트 #SoloStack [en] GitHub Trending Top 10 Developers Shouldn't Miss This Week (Week 2, October) This week's (10/7–10/14) GitHub Trending weekly Top 10, plus the hottest repo outside the ranking. The order is exactly as on GitHub's Trending (weekly) page (collected 2026-10-14 06:00 KST). Star counts are as of collection. 1. AnyPS5 — relinks PS5 executables for Linux and Windows without an emulator (GPL-2.0). The game footage in the video (Animal Well, Ape Academy 2, Astro Bot, Balatro, Dreaming Sarah, Hello Neighbor 2) is the Windows run footage posted in the repo README. ⚠ The docs don't say how to obtain game executables, and legality is left to the user. Concerns about bypassing copy protection (DMCA, etc.) have been raised, so it's covered only briefly. 2. mattpocock/skills — coding agent skills that work in the order: ask back → spec → tickets → TDD → review (MIT) 3. tester-army/e2e — E2E tests mixing in natural-language steps (0.19.0 toHaveScreenshot, 0.20.0 run --watch) 4. EpicGames/raddebugger — a native debugger that handles multiple processes together (ALPHA, preliminary Linux support) 5. heygen-com/hyperframes — write HTML, render to video (transcribe --to srt; fixed a path where turning off telemetry was ignored, v0.8.151) 6. thedotmack/claude-mem — memory across coding agent sessions (passed 100K stars; from v13.37.0, falls back to your Anthropic plan by default when the provider fails) 7. cursor/plugins — Cursor's plugin spec and official plugins (Workday, Salesforce and Square integrations added) 8. earthtojake/text-to-cad — design 3D parts in words (telemetry on by default since 10/8) 9. flutter/flutter — 3.47.7 hotfix: two iOS accessibility bugs fixed 10. cathrynlavery/diagram-design — ask in words, get an architecture diagram as a single HTML+SVG file (45 kinds) Outside the ranking: morluto/rea — connects agents to reverse engineering tools (stars 8,692 → 203,368, 10/7 → 10/14 via GitHub API · 797 points on HN 10/10 · not on the weekly trending page) ▶ Turning off telemetry (usage reporting) - text-to-cad: uvx cadgen telemetry off or DO_NOT_TRACK=1 - claude-mem: npx claude-mem telemetry disable · block the automatic plan fallback with CLAUDE_MEM_QUOTA_FALLBACK_PROVIDER=off - hyperframes: HYPERFRAMES_NO_TELEMETRY=1 (update to v0.8.151 or later) - e2e: E2E_TELEMETRY_DISABLED=1 - Flutter: flutter --disable-analytics ▶ Also worth a look this week - vercel-labs/skills (npx skills) v1.7.2: install and sync team-shared skills via a skills field in package.json. Telemetry on by default (DISABLE_TELEMETRY=1) - debpalash/VoiceStudio: remote devices and workers moved to paid Pro (10/11); the default voice model with unclear commercial licensing now requires consent before use (10/10) ▶ Related videos - REA, tried it myself: youtu.be - e2e measured: youtu.be ▶ Links boykopovar/AnyPS5 mattpocock/skills tester-army/e2e EpicGames/raddebugger heygen-com/hyperframes thedotmack/claude-mem cursor/plugins earthtojake/text-to-cad flutter/flutter cathrynlavery/diagram-design morluto/rea Narration is an AI voice (ElevenLabs). For repos without real footage (text-to-cad, RAD Debugger, mattpocock/skills, AnyPS5), the video shows explanatory motion graphics of the steps described in their official docs. #GitHubTrending #OpenSource #Developers #AIagents #CodingAgents #SoloStack [en-US] Top 10 GitHub Trending Repositories Developers Shouldn't Miss This Week (2nd Week of Oct) I have compiled the Top 10 GitHub Trending Weekly rankings for this week (October 7–14), along with one repository that was the hottest outside the rankings. The rankings follow the exact order on the GitHub Trending (weekly) page (collected on October 14, 2026, 06:00 KST). The star ratings are based on the values ​​at the time of collection. 1st Place: AnyPS5 — A tool that relinks PS5 executable files for Linux and Windows without an emulator (GPL-2.0). The game screenshots in the video (Animal Well, Ape Academy 2, Astro Bot, Balatro, Dreaming Sarah, Hello Neighbor 2) are Windows execution screenshots uploaded to the repository's README. ⚠ The documentation does not explain how to obtain game executable files, and legality is left to the user. I have only briefly touched upon this topic due to concerns regarding copy protection bypasses (such as DMCA). 2nd place mattpocock/skills — A collection of coding agent skills that make it work in the order of Rephrasing → Spec → Ticket → TDD → ​​Review (MIT) 3rd place tester-army/e2e — E2E testing that mixes natural language stages (0.19.0 toHaveScreenshot, 0.20.0 run --watch) 4th place EpicGames/raddebugger — Native debugger that handles multiple processes together (ALPHA, Linux pre-support) 5th place heygen-com/hyperframes — Write in HTML and render as video (transcribe --to srt, fixed path to ignore telemetry off v0.8.151) 6th place thedotmack/claude-mem — Remember between coding agent sessions (Reached 100,000 stars, defaults to user's Anthropic plan in case of provider failure starting v13.37.0) 7th place cursor/plugins — Cursor plugin specification and official plugin (Workday, Salesforce, Square integration added) 8th place earthtojake/text-to-cad — Design 3D parts by voice (Telemetry enabled by default from 10/8) #9 flutter/flutter — 3.47.7 Hotfix: Fixed 2 iOS accessibility bugs #10 cathrynlavery/diagram-design — Creates diagrams as a single HTML+SVG file via voice commands (45 types) #Out of Rank morluto/rea — Connect agents to reverse engineering tools (Stars 8,692 → 203,368, 10/7 → 10/14 GitHub API · HN 797 points 10/10 · Not on Weekly Trending Page) ▶ How to turn off Telemetry (transmit usage history) - text-to-cad: uvx cadgen telemetry off or DO_NOT_TRACK=1 - claude-mem: npx claude-mem telemetry disable · Prevent automatic plan switching CLAUDE_MEM_QUOTA_FALLBACK_PROVIDER=off - hyperframes: HYPERFRAMES_NO_TELEMETRY=1 (Updated to v0.8.151 or higher) - e2e: E2E_TELEMETRY_DISABLED=1 - Flutter: flutter --disable-analytics ▶ Things to look at this week - vercel-labs/skills(npx skills) v1.7.2: Install and clean up team shared skills to the skills field in package.json (sync). Telemetry On by default (DISABLE_TELEMETRY=1) - debpalash/VoiceStudio: Remote device and worker features moved to paid Pro (10/11), default voice models with unclear commercial licenses require consent (10/10) ▶ Related Videos - Video of REA personal testing: youtu.be - e2e actual measurements: youtu.be ▶ Links boykopovar/AnyPS5 mattpocock/skills tester-army/e2e EpicGames/raddebugger heygen-com/hyperframes thedotmack/claude-mem cursor/plugins earthtojake/text-to-cad flutter/flutter cathrynlavery/diagram-design morluto/rea ※ The narration is an AI voice (ElevenLabs). Repositories without actual footage (text-to-cad, RAD Debugger, mattpocock/skills, AnyPS5) are principle motions that draw the sequence of actions described in the official documentation. #GitHubTrending #OpenSource #Developer #AIAgent #CodingAgent #SoloStack