msitarzewski

1. Agency Agents (comunidad) +160.000 estrellas. Unos 170 agentes organizados como una empresa: ingeniería, diseño, marketing, ventas, producto, testing, seguridad y soporte. Compatible con Claude Code, Cursor y más. msitarzewski/agency-agents

Install a team of specialist agents into Claude Code Generic prompts give generic answers, and writing a good persona for every role takes hours. Agency Agents is a collection of ready-made specialist agent files (frontend developer, backend architect, DevOps automator and many more) that you can drop into Claude Code. Below is the route from the README, with the real commands. The short version is on the Deduped use-case page. What you get Each agent file holds an identity and personality, a core mission and workflow, technical deliverables with code examples, and success metrics. The idea is that you activate the specialist that fits the task instead of re-explaining a role every session. Fastest install The README now recommends a native desktop app for macOS, Linux and Windows that browses the roster and installs agents into Claude Code, Cursor, Codex, Gemini CLI and others. On a Mac: brew install --cask msitarzewski/agency-agents/agency-agents Prefer the terminal? The scripts install the same agents. Install with the script Clone the repo, then list what exists: ./scripts/install.sh --list teams Install only the divisions you need into Claude Code: ./scripts/install.sh --tool claude-code --division engineering,security Or install everything for Claude Code: ./scripts/install.sh --tool claude-code Or copy one division by hand cp engineering/*.md ~/.claude/agents/ That is all Claude Code needs: the files sit in its agents directory. Use a specialist Start a Claude Code session and ask for a role by name. The README's own example: > Hey Claude, activate Frontend Developer mode and help me build a React component For a feature that spans layers, split the work: > Use the Backend Architect agent to design the API for <feature>, then the Frontend Developer agent to build its UI. Summarize what each agent produced. Switch to the DevOps Automator agent when the task becomes pipelines and deployment. Other tools The same repo can generate integration files for other tools. The README lists GitHub Copilot, Gemini CLI, OpenCode, Cursor, Aider, Windsurf, Codex and more. Run the converter, then install: ./scripts/convert.sh ./scripts/install.sh --tool gemini-cli The Deduped page on using these agents in Cursor or Codex covers that route. Things to know - Installing every division loads a lot of agent files. Start with engineering and security, and add more only when you use them. - A persona does not replace review. Read what the Backend Architect proposes the way you would read a colleague's design. - Agent files are prompts that shape behavior, so skim one before installing a collection from any repo. Setup takes about 10 minutes: one install command, one session, one named specialist. If you want to trim what your agent loads, compare with the Deduped page on installing a startup MVP runbook team. Want the setup commands for 50 AI repos people actually ship with? Get the free AI Repos That Actually Ship pack: dedupedai.com

agency-agents drops ~300 specialist markdown files (engineering, security, design) that install into Claude Code or Cursor via one shell script. does the agent load only the activated specialist or keep the full roster in context on every turn? msitarzewski/agency-agents

I went through all 282 agents in that agency-agents repo. Interesting but not yet scary. Hi - Ive spent 20 years working in and consulting with roughkly 200 agencies, mostly 30-90 people. Lots of talk there about what people to keep, etc. I saw the GitHub repo mention (on LI) went down the rabbit hole. This note was drafted by Claude from a larger article drafted by Claude, based on what we did. We went through all 282 files in github.com/msitarzewski/agency-agents. That’s 537,066 words. Could’ve found a better use for a weekend. But it started getting cool as I combined datasets. First, the basics of the repo. You can pretty well tell who built it from the org chart. There are 85 engineering agents, 27 for games and VR, 13 for GIS and mapping. Account and client services gets four. PR gets one. So there’s that. The agents get pretty detailed where there’s already a rulebook. Healthcare marketing compliance gets 4,638 words. Legal billing gets 3,585. Content Creator? 335 words. No rules, no workflow. Same deal with rules and workflow for all seven paid media agents. And when you get to the judgment calls, you find a lot of blanks. Judgment calls are what an agency (really) gets paid for. For example, the account strategist agent has a QBR (quarterly business review, a standard account-person meeting) prep includes this: “[Who should be there but isn’t — and why]” And that IS a great question. But how would the agent know? That answer comes from knowing the client and what’s been going on over there. Not just Human IN the Loop, but Human IS the Loop. We counted about 465 blanks like that. Another 93 agents have a “Learning & Memory” section that says “Remember and build expertise in:” followed by empty headings. Nothing in the repo actually stores what they learn. There's more, but that's the shape of coverage, and it is kind of what I would expect given that it was probably built mostly by technologists, That many of the other roles that are very scant are the ones where a lot of situated judgment is used… Which AI isn't very good at. One other question that came up at this point was regarding how thorough the agents were in terms of the topics and things that needed to deal with. I had Claude compare them with 423 actual agency job postings across 15 roles, junior through senior, and a few other data sets. We sorted the responsibilities into procedural work and stuff that depends on the client. The agents were 71% procedural focused. Real jobs were 33%. Even junior jobs were only 47%. So kind of not ready for prime time. Probably worth exploring but what they are missing may be pretty costly to suss out, versus getting more narrow and building your own. One of my former clients and a really good friend, Bret Starr, has been hacking away at this problem for three years now, and he's been able to build a credible agent swarm that I think number is something like 30 to 40 agents, that can do most of what needs to be done in a narrow B2B marketing niche. So this kind of stuff can be done, but again, this has been him dedicating most of his life cycles to building this thing for three years. I also built a modeling tool so that I can start playing with modeling staffing based on how much work AI can take over in the divide between assembly work versus judgment work… Happy to share that if anyone's interested. Not sure it really applies well at the smaller scale like five or 10 person agency. It shows that it is reasonable to get a 40% reduction in junior staff and something less than 15% in mid and none in senior, based on today's capabilities. Anyway, I spent too much time on this. Full write-up here if you want the rest. Feedback welcome and always up for a quick chat if there something more I can share. I have another more detailed analysis that I'm still working on, so lmk if I should post it here. (not astroturfing (cool term) advertising or anything else. I'm sure the mods will let me know if I broke a rule)

4. agency-agents 这周新增 3,068 颗 Star，总共约 16 万颗。 相当于拥有一整家 AI 公司的不同岗位角色： 营销、社交媒体运营、商务提案、广告投放…… 需要哪个角色，就可以调用哪个。 它还有自己的桌面 App，点击一下就能安装到 Claude Code 中。 GitHub：msitarzewski/agency-agents

GitHub Repo Goldmine #01: agency-agents Agency-agents has 164K GitHub stars (as of 27 Oct 2026) and 200+ free AI agent personas for Claude Code, Cursor, GitHub Copilot, Codex, Gemini CLI and more. Each one is just a Markdown file. The 2 gems worth installing: 💎 Minimal Change Engineer: the smallest possible fix, no scope creep, every changed line has to be justified. 💎 Reality Checker: its default verdict is "NEEDS WORK" until it sees screenshots and test results. Install only the teams you need: git clone msitarzewski/agency-agents cd agency-agents ./scripts/install.sh --tool claude-code --division engineering,testing They're expert prompts, not employees. Always review what they produce. Sources: • Repo: msitarzewski/agency-agents • Reality Checker: github.com/msitarzewski/agency-agents/blob/main/t… • Minimal Change Engineer: github.com/msitarzewski/agency-agents/blob/main/e… • Install guide: flaviocopes.com 🤖 ByteBuddy digs up a new GitHub repo every week. Subscribe for the next Goldmine. #GitHub #AIAgents #ClaudeCode #OpenSource #Shorts

Dieses GitHub-Projekt gibt dir ein komplettes KI-Team Entwickler, Designer, Marketing-Profi: Dieses GitHub-Projekt liefert dir die vollständigen Profile für dein eigenes KI-Team. Es heißt Agency Agents, hat über 160.000 Sterne und bringt 283 fertige Rollen in 18 Abteilungen mit. So sieht das in der Praxis aus: Für die Oberfläche deiner App holst du dir den Frontend Developer, für den Launch den Growth Hacker, und am Ende prüft der Reality Checker alles kritisch (sein Standard-Urteil: „NEEDS WORK“). Jede Rolle bringt Anweisungen, Arbeitsabläufe und klare Vorgaben mit. Du passt sie an dein Projekt an und nutzt sie zum Beispiel in Claude Code oder Cursor. Kostenlos, MIT-Lizenz. 🔗 Agency Agents auf GitHub: msitarzewski/agency-agents 📄 Anleitung: ki-community.worthmaking.de #KIAgenten #ClaudeCode #Cursor #GitHub #OpenSource #KI #VibeCoding #AgencyAgents #Shorts

7 kostenlose Repos, die Claude unglaublich mächtig machen | GitHub | KI-Skills | Konnektoren Claude hat mittlerweile tausende Repos – aber diese 7 Claude Skills, Tools und Konnektoren können deinen kompletten Workflow verändern. Von echter Marktrecherche über Kundenfindung und SEO bis hin zu Videos, Werbekampagnen und einem Team aus 282 KI-Agenten: Ich teste alle sieben an einem einzigen Beispiel. 🔗 REPOS: last30days: mvanhorn/last30days-skill gstack: garrytan/gstack Apify MCP Server: apify/apify-mcp-server OpenSEO: every-app/open-seo HyperFrames: heygen-com/hyperframes Claude Ads: AgriciDaniel/claude-ads agency agents: msitarzewski/agency-agents Das erwartet dich im Video: ✅ last30days – findet heraus, worüber echte Menschen gerade sprechen ✅ gstack – prüft deine Geschäftsidee wie ein Startup-Investor ✅ Apify – findet potenzielle Kunden und relevante Unternehmen ✅ OpenSEO – analysiert Keywords, Rankings und KI-Sichtbarkeit ✅ HyperFrames – erstellt Videos direkt aus Claude ✅ Claude Ads – entwickelt und prüft Werbekampagnen ✅ agency agents – gibt dir Zugriff auf 282 spezialisierte KI-Rollen Das Spannende daran: Ich teste die Tools nicht einzeln, sondern verbinde sie zu einem kompletten Workflow – von der ersten Idee über Recherche und Validierung bis zur fertigen Marketingkampagne. Kapitel: 00:00 Intro – 7 kostenlose Claude Repos, Skills & Konnektoren 00:23 last30days – Marktrecherche mit echten Meinungen 01:50 gstack – Geschäftsideen kritisch prüfen 03:33 Apify – Potenzielle Kunden finden 04:51 OpenSEO – SEO, Keywords & KI-Sichtbarkeit 05:59 HyperFrames – KI-Videos erstellen 06:52 Claude Ads – Werbekampagnen planen & optimieren 08:03 agency agents – 282 KI-Spezialisten für Claude 09:26 Fazit – Welche Tools lohnen sich wirklich? **LASS UNS VERNETZEN** ✔️ Website: shakeelazq.de ✔️ Instagram: instagram.com ✔️ TikTok: tiktok.com ✔️ LinkedIn: linkedin.com ✔️ Newsletter: 8g2tqd.subscribepage.io **✉️ BUSINESS-ANFRAGEN** Für Kooperationen, Projekte und geschäftliche Anfragen erreichst du mich hier: kontakt.shakeelazq@gmail.com #Claude #ClaudeAI #GitHub #ClaudeSkills #KI

First, you can setup Agency Agents inside Claude Code from here: msitarzewski/agency-agents Second: If you want to learn how to build systems & then apply them to real businesses to get clients, check out Maker School. It's my AI agent & automation community, where I teach you how to build Claude Code, Codex, & n8n based flows, and sell them to your first client in 90 days. Includes a 100% money back guarantee if you don't get a paying client! skool.com

깃허브에 무료로 풀린 AI 직원 665명, 부서별로 골라 쓰세요 깃허브에 AI 직원 665명이 무료로 풀렸습니다. 마케팅·사장단·제품·영업·보안… 부서별로 골라 쓰면 되고, 일하는 순서·체크리스트·결과물 양식까지 미리 들고 옵니다. 클로드 코드·코덱스·제미나이·커서에서 쓸 수 있어요. · 출처 1: alirezarezvani/claude-skills (MIT, 깃허브 별 28,582) — alirezarezvani/claude-skills · 출처 2: msitarzewski/agency-agents (MIT, 깃허브 별 163,403) — msitarzewski/agency-agents · 665 = 두 모음을 합쳐 같은 이름은 한 번만 센 수(10-23 기준). 별은 합쳐 19만 개 넘음. · 저장소는 무료(MIT, 회사 일에도 OK)이고, 쓰는 AI 도구 요금은 따로예요. 좋아요 누르고 댓글에 「직원」 남겨 주세요. 명단이랑 설치 링크는 고정 댓글에 있어요. #AI #AI직원 #클로드코드 #깃허브 #업무자동화 [de-DE] 665 kostenlose KI-Mitarbeiter auf GitHub: Wählen Sie sie nach Abteilungen aus 665 KI-Mitarbeiter wurden kostenlos auf GitHub veröffentlicht. Sie können aus Abteilungen wie Marketing, Führungskräfte, Produktmanagement, Vertrieb und Sicherheit wählen. Die Mitarbeiter sind mit Arbeitsabläufen, Checklisten und vorgegebenen Lieferformaten ausgestattet und können in Claude Code, Codex, Gemini und Cursor verwendet werden. • Quelle 1: alirezarezvani/claude-skills (MIT-Lizenz, 28.582 GitHub-Sterne) — alirezarezvani/claude-skills • Quelle 2: msitarzewski/agency-agents (MIT-Lizenz, 163.403 GitHub-Sterne) — msitarzewski/agency-agents • 665 = Anzahl der Kombinationen zweier Vokale, wobei gleiche Namen nur einmal gezählt werden (basierend auf dem Zahlensystem 10–23). Die Gesamtzahl der Sterne übersteigt 190.000. Das Repository ist kostenlos (MIT-Lizenz, für die Nutzung im Unternehmen geeignet), die Kosten für die verwendeten KI-Tools fallen jedoch separat an. Bitte liken und „Employee“ in die Kommentare schreiben. Die Liste und der Installationslink befinden sich im angepinnten Kommentar. #KI #AIEmployee #ClaudeCode #GitHub #Arbeitsautomatisierung [es-US] 665 empleados de IA gratis en GitHub: elíjalos según el departamento Se han publicado gratuitamente 665 empleados de IA en GitHub. Puedes elegir entre departamentos como Marketing, Ejecutivos, Producto, Ventas y Seguridad, y vienen precargados con secuencias de trabajo, listas de verificación y formatos de entrega. Se pueden usar en Claude Code, Codex, Gemini y Cursor. · Fuente 1: alirezarezvani/claude-skills (MIT, 28.582 estrellas en GitHub) — alirezarezvani/claude-skills · Fuente 2: msitarzewski/agency-agents (MIT, 163.403 estrellas en GitHub) — msitarzewski/agency-agents · 665 = El número de veces que se combinan dos vocales y el mismo nombre se cuenta solo una vez (basado en 10-23). ​​El número total de estrellas supera las 190.000. • El repositorio es gratuito (licencia MIT, apto para uso empresarial), pero las herramientas de IA utilizadas tienen un coste adicional. Dale a "Me gusta" y escribe "Employee" en los comentarios. La lista y el enlace de instalación se encuentran en el comentario fijado. #IA #EmpleadoIA #ClaudeCode #GitHub #AutomatizaciónDelTrabajo [fr-FR] 665 employés IA disponibles gratuitement sur GitHub : choisissez-les selon vos départements 665 employés virtuels sont disponibles gratuitement sur GitHub. Vous pouvez choisir parmi des services tels que le marketing, la direction, le produit, les ventes et la sécurité. Ces employés sont préconfigurés avec des séquences de tâches, des listes de contrôle et des formats de livrables. Ils sont compatibles avec Claude Code, Codex, Gemini et Cursor. • Source 1 : alirezarezvani/claude-skills (MIT, 28 582 étoiles sur GitHub) — alirezarezvani/claude-skills • Source 2 : msitarzewski/agency-agents (MIT, 163 403 étoiles sur GitHub) — msitarzewski/agency-agents • 665 correspond au nombre de combinaisons de deux voyelles, chaque nom identique n’apparaissant qu’une seule fois (sur la base de 10 à 23). Le nombre total d’étoiles dépasse 190 000. Le dépôt est gratuit (licence MIT, utilisable en entreprise), mais les frais des outils d'IA utilisés sont à votre charge. N'hésitez pas à aimer et à laisser un commentaire avec « Employé ». La liste des outils et le lien d'installation se trouvent dans le commentaire épinglé. #IA #IAEmployé #ClaudeCode #GitHub #AutomatisationDuTravail [hi] गिटहब पर मुफ्त में उपलब्ध 665 एआई कर्मचारी, विभाग के अनुसार चुनें और उपयोग करें GitHub पर 665 AI कर्मचारी मुफ्त में उपलब्ध हैं। आप मार्केटिंग, एग्जीक्यूटिव, प्रोडक्ट, सेल्स और सिक्योरिटी जैसे विभागों में से चुन सकते हैं। ये कर्मचारी पहले से ही कार्य क्रम, चेकलिस्ट और डिलिवरेबल फॉर्मेट के साथ आते हैं। इनका उपयोग Claude Code, Codex, Gemini और Cursor में किया जा सकता है। · स्रोत 1: alirezarezvani/claude-skills (MIT, 28,582 GitHub स्टार) — alirezarezvani/claude-skills · स्रोत 2: msitarzewski/agency-agents (MIT, 163,403 GitHub स्टार) — msitarzewski/agency-agents · 665 = दो स्वरों के संयोजन की संख्या, जिसमें एक ही नाम को केवल एक बार गिना जाता है (10-23 के आधार पर)। कुल स्टारों की संख्या 190,000 से अधिक है। · यह रिपॉजिटरी मुफ़्त है (MIT लाइसेंस के तहत, कंपनी के काम के लिए ठीक है), लेकिन इस्तेमाल किए गए AI टूल्स के लिए शुल्क अलग से देना होगा। कृपया लाइक करें और कमेंट में "Employee" लिखें। लिस्ट और इंस्टॉलेशन लिंक पिन किए गए कमेंट में हैं। #AI #AIEmployee #ClaudeCode #GitHub #WorkAutomation [id] 665 AI pekerja gratis di GitHub, pilih dan gunakan sesuai departemen 665 karyawan AI telah dirilis secara gratis di GitHub. Anda dapat memilih dari departemen seperti Pemasaran, Eksekutif, Produk, Penjualan, dan Keamanan, dan mereka sudah dilengkapi dengan urutan kerja, daftar periksa, dan format hasil kerja. Mereka dapat digunakan di Claude Code, Codex, Gemini, dan Cursor. · Sumber 1: alirezarezvani/claude-skills (MIT, 28.582 bintang GitHub) — alirezarezvani/claude-skills · Sumber 2: msitarzewski/agency-agents (MIT, 163.403 bintang GitHub) — msitarzewski/agency-agents · 665 = Jumlah kali dua huruf vokal digabungkan dan nama yang sama hanya dihitung sekali (berdasarkan 10-23). ​​Jumlah total bintang melebihi 190.000. • Repositori ini gratis (MIT, cocok untuk pekerjaan perusahaan), tetapi biaya untuk alat AI yang digunakan terpisah. Silakan beri like dan tinggalkan "Employee" di kolom komentar. Daftar dan tautan instalasi ada di komentar yang disematkan. #AI #AIEmployee #ClaudeCode #GitHub #WorkAutomation [it] 665 impiegati AI rilasciati gratuitamente su GitHub, sceglieteli per dipartimento Sono stati rilasciati gratuitamente su GitHub 665 dipendenti IA. È possibile scegliere tra reparti come Marketing, Dirigenza, Prodotto, Vendite e Sicurezza, e sono precaricati con sequenze di lavoro, checklist e formati di consegna. Possono essere utilizzati con Claude Code, Codex, Gemini e Cursor. · Fonte 1: alirezarezvani/claude-skills (MIT, 28.582 stelle su GitHub) — alirezarezvani/claude-skills · Fonte 2: msitarzewski/agency-agents (MIT, 163.403 stelle su GitHub) — msitarzewski/agency-agents · 665 = Il numero di volte in cui due vocali vengono combinate e lo stesso nome viene contato una sola volta (basato sulla notazione 10-23). Il numero totale di stelle supera le 190.000. Il repository è gratuito (licenza MIT, utilizzabile per scopi aziendali), ma i costi per gli strumenti di IA utilizzati sono a parte. Metti "Mi piace" e scrivi "Employee" nei commenti. L'elenco e il link per l'installazione si trovano nel commento fissato in alto. #AI #AIEmployee #ClaudeCode #GitHub #WorkAutomation [pt-BR] 665 funcionários de IA liberados gratuitamente no GitHub, escolha por departamento 665 funcionários de IA foram disponibilizados gratuitamente no GitHub. Você pode escolher entre departamentos como Marketing, Executivos, Produto, Vendas e Segurança, e eles vêm pré-carregados com sequências de trabalho, listas de verificação e formatos de entregáveis. Podem ser usados ​​no Claude Code, Codex, Gemini e Cursor. · Fonte 1: alirezarezvani/claude-skills (MIT, 28.582 estrelas no GitHub) — alirezarezvani/claude-skills · Fonte 2: msitarzewski/agency-agents (MIT, 163.403 estrelas no GitHub) — msitarzewski/agency-agents · 665 = Número de vezes que duas vogais são combinadas e o mesmo nome é contado apenas uma vez (baseado em 10-23). ​​O número total de estrelas ultrapassa 190.000. O repositório é gratuito (MIT, adequado para uso corporativo), mas as taxas das ferramentas de IA utilizadas são cobradas separadamente. Curta e comente "Funcionário". A lista e o link de instalação estão no comentário fixado. #IA #FuncionárioIA #ClaudeCode #GitHub #AutomaçãoDeTrabalho