Claude Skills by smk-labs

Write Persian web copy (blog articles, product descriptions, landing pages, category text, microcopy, meta titles and descriptions) using Cursor workers, so the writing runs on the Cursor subscription's quota instead of Claude's. Use when the user asks for Persian content for a site, says "متن فارسی بنویس", "مقاله بنویس", "توضیح محصول بنویس", "کپی‌رایتینگ فارسی", "محتوای سئو", "copy for the site in Persian", or wants blog/product/landing copy generated at volume. Also use to edit, proofread, ...

Delegate a coding or research task to the Cursor CLI (cursor-agent) so it runs on the Cursor subscription's quota instead of Claude's, with Claude still orchestrating. Use when the user says "delegate to cursor", "run this with cursor-agent", "offload to cursor", "spend the cursor quota", "have cursor do it", or wants to hand a heavy self-contained slice of a larger job to a second agent. Works with the user's default Cursor login; an optional account name targets a specific seat.

Orchestrate a fleet of Cursor CLI (cursor-agent) workers to execute a large or multi-part job, while Claude does the architecture, UX/UI decisions, decomposition, review, and integration. Use when the user says "orchestrate with cursor", "build this with a cursor fleet", "fan out to cursor agents", "cursor workflow", "have cursor build X", or when a delegation is too big for one worker. Optional mode word: "workflow"/"w" forces the JS harness; "sub"/"subagent" forces a single simple worker. O...

وضعیت فوق‌خلاصه + ادامهٔ خودکار کار ناتمام. وقتی سشن وسط کار قطع شده یا کاربر می‌گوید «ادامه», «ادامه بده», «وضعیت چیه», «چه خبر», «go», «/go», or wants to resume after a disconnect: سه سطر وضعیت بده و بلافاصله از همان نقطه ادامه بده. Never restart from scratch.

فقط برای زبان فارسی. هر جوابی که به فارسی نوشته می‌شود باید ساده، کوتاه، مرتب و آدم‌وار باشد؛ نه ترجمه‌زده، نه اداری، نه پرحرف. همیشه روی متن فارسی اعمال شود. Persian-only writing style: plain, concise, well-structured Persian. Does not apply to English or any other language.

تحقیق سبک و سریع چندایجنتی (mini research؛ عمداً نه deep research) روی یک موضوع. ورودی‌ها: موضوع، و اختیاری: مدل ایجنت‌ها (haiku یا sonnet یا opus)، نکات ویژهٔ تمرکز، تعداد ایجنت، ذخیرهٔ خروجی. Use when the user says "mini-research", "/mini-research", "تحقیق سبک", "یه ریسرچ کوچیک", "چند ایجنت بگرد", "با چند ایجنت هایکو بگرد", "با ایجنت‌های اپوس بگرد", "mini research on X".

پرامپت‌ساز تشریحی: پرامپت کامل با نقش، کانتکست، شرح کار، گاردریل و فرمت گزارش، برای وقتی محیط عجیب است یا قید سفت دارد (دسترسی، خط قرمز، لحن، فرمت خروجی دقیق). خروجی هم پیشنهاد تسک یک‌کلیکی است و هم متن خام در چت. Use when the user says "پرامپت کامل", "پرامپت تشریحی", "با گاردریل", "پرامپت با نقش و کانتکست", "spec", "classic", "/mkprompt-classic", or gives hard constraints the agent must follow exactly. برای پرامپت معمولی و هدف‌محور، به‌جای این از mkprompt-goal استفاده کن.

پرامپت‌ساز هدف‌محور و سبک پیش‌فرض: به‌جای دیکته کردن مراحل، هدف را یک سطح بالاتر می‌گوید، به مدل راه تأیید کار خودش را می‌دهد و شرط توقف عینی می‌گذارد؛ برای کار چندسشنی خودِ پرامپت لوپ و حافظه و ساب‌ایجنت‌هایش را می‌سازد. خروجی هم پیشنهاد تسک یک‌کلیکی است و هم متن خام در چت. Use when the user says "پرامپت بساز", "یه پرامپت آماده کن برای", "پرامپت بده", "پرامپت هدف‌محور", "پرامپت ساده", "mkprompt", "/mkprompt-goal", "write me a prompt for", or asks to prepare a standalone prompt to hand to ano...

پرامپت‌ساز بازطراحی: پرامپتی می‌سازد که یک طراحی موجود را در چندین راند واقعی جلو ببرد و نگذارد راندها به رنگ و فاصله تنزل کنند. هر راند اول جواب می‌دهد «اگر این وجود نداشت، برای آدمی که این کار را می‌کند چه می‌ساختیم؟» و بعد تصمیم می‌گیرد چه عوض شود؛ حذف صفحه، ادغام فلو و برداشتن مرحله مجاز است. خروجی هم پیشنهاد تسک یک‌کلیکی است و هم متن خام در چت. Use when the user says "پرامپت بازطراحی", "پرامپت ریدیزاین", "پرامپت بده این رو بازطراحی کنه", "چند راند روی این طراحی کار کن", "این طراحی رو بهت...

گزارش را با readable بساز، بعد از تأیید روی پورتال صفحه‌اش کن، و آخرش یک ایمیل کوتاه به لحن خودم با لینک همان صفحه بده. Use when the user says "منتشر کن", "بذار روی پورتال", "صفحه‌اش کن", "لینکش کن", "publish this", "put it on the portal", "share this as a page", "/publish".

دستورالعمل گام‌به‌گام به زبان ساده برای انجام یک کار: گرفتن کلید API، راه‌اندازی یک سرویس یا ابزار، کار اداری، هر روال چندمرحله‌ای. فقط گام‌هایی که خود کاربر باید انجام دهد. Use when the user says "چطور", "چه جوری", "دستورالعمل بده", "گام به گام بگو", "steps", "/steps", "how do I get/set up X", "step by step instructions".

کارزار کامل «سوارم کیوای» روی یک محصول زنده با متد چهار فازی: ممیزی استاتیک سوارمی کد، کیوای عمیق زنده، سوارم فیکس، و کیوای تأیید نهایی. با چک‌پوینت روی دیسک و گزارش تاریخ‌دار. Use when the user says "سوارم کیوای", "سوارم", "swarm qa", "/swarm-qa", "کیوای کامل بزن", or asks for a full multi-agent QA campaign over a live stack. Not for a quick single-page check.

کدبیس رو مثل یه سینیور برای مدیر محصول قصه می‌کنه: قصهٔ دو دقیقه‌ای از بالا، درخت فایل قابل‌گشتن مثل VS Code با وزن‌دهی اهمیت، کانواس خلاصه برای هر فایل، درس‌های کوچیک مهندسی برای مسیر PM به دولوپر، و ترجمهٔ دیف یا PR مثل ریلیز نوت. Use when the user says «دیکد», «decode», «/decode», «قصهٔ کدبیس», «این فایل چی کار می‌کنه», «این کد رو به زبان ساده توضیح بده», «برای پی‌ام توضیح بده», «به زبان محصول», «اطلس کدبیس», «نقشهٔ کدبیس», «این تغییرات یعنی چی», "explain this file for a PM", "what does th...

Make a fig. A single looping animated SVG, one self-contained HTML file you can drop in an email or a slide. Use it when an idea moves (flows, loops, retries, queues, fan-outs). Faster than a paragraph, livelier than a static diagram. No player, no deck.

Search and fetch free, licensed stock images from the internet (Pexels, Pixabay, Unsplash, Openverse, Wikimedia Commons). Two jobs: add optimized WebP images to sites, blogs, and docs, or show a fitting photo right in the chat while answering about a topic. Use when the user wants to find, download, insert, or see a real photo, e.g. "find an image of X", "show me a photo of X", "add a hero photo", "getpix", or Persian requests like "عکس پیدا کن" / "عکسشو نشونم بده" / "عکس بذار". Also use proa...

Run a conservative, end-to-end art-direction photo pass over a whole website: audit where photos genuinely earn their place, define one visual language, source free images, melt them into the design, verify in both themes, and record credits. Use when the user asks for a "photo pass", wants photography/warmth/imagery added across a site or landing page ("add photos to the site", "make it feel warmer with real photos"), or Persian requests like "به سایت عکس بده" / "پاس عکس بزن". For a single o...

The way into learnable and the way back. If a course already exists here it continues from exactly where the learner stopped, next lesson, next chapter, no re-planning. If nothing exists yet it asks whether to measure their level first or go straight to building a course. Use when the user says they want to learn, study, continue, resume, keep going, "ادامه", "درس بعدی", "بریم یاد بگیریم", "شروع کنیم", "continue the course", "next lesson", "where were we", or invokes /learnable:learn. Also th...

Write the next book of the course: its chapters and every lesson inside them, ready to teach. Reads the plan, runs a short research pass only for chapters that need one, then writes one file per chapter with the opening question, the correction, the explanation, the terms with their origins, and the end-of-chapter quiz. Use when the learner has finished a book, when a chapter file is missing, or when the user says "کتاب بعدی", "درس‌های بعدی رو بنویس", "فاز بعد", "write the next book", "author...

Build a whole course on a subject, from scratch, on disk. A light survey of the domain, then the placement-quiz decision, then a deep multi-agent research pass that becomes an ordered taxonomy, a table of contents grouped into books and chapters, and a course folder that teaches itself. Use when the user asks for a course, curriculum, syllabus, study plan, roadmap or taxonomy on a topic, says "دوره بساز", "کورس جدید", "برنامه درسی", "نقشه یادگیری", "build me a course on X", "I want to learn X...

Placement quiz that finds someone's real level in a subject and writes it to PROFILE.md. Scenario questions in small batches, never definitions, with neutral options that cannot be eliminated by tone, plus self-report checks that separate knowing a word from being able to act on it. Use when the user asks to measure their level, wants a placement, diagnostic or self assessment, says "تعیین سطح", "آزمون بگیر", "سطحم رو بسنج", "خودارزیابی", "test my level", "where am I in X", "assess me on X", ...

Use the Open Design content library: 151 design systems with real tokens (Apple, Vercel, Linear, Stripe, Notion, GitHub, Raycast, plus styles like brutalism and claymorphism), 71 design and frontend skills, and 114 rendering templates for decks, documents, video frames, and social cards. Use when building or restyling any UI, picking a visual direction, needing real design tokens instead of invented ones, matching a known brand's look, animating with GSAP, or producing a deck, report, poster,...

Generate a GitLab CI/CD pipeline and Helm chart that builds the current project with kaniko, pushes to the GitLab Container Registry, and deploys to a Kubernetes namespace via a kubeconfig CI variable. Use when the user asks to set up CI/CD, create a GitLab pipeline, write a Helm chart, containerize and deploy an app to Kubernetes/GKE, or wire up "push to deploy". Also generates a Dockerfile tailored to the detected runtime when one is missing, handles backend + frontend split repos, and cove...

Right-size Kubernetes pod CPU/memory requests from real Prometheus usage data (per-pod p95 CPU, p98 memory over 7 days), compute how many nodes the cluster needs, and tune KEDA queue-based autoscaling. Use when the user asks to rightsize workloads, fix over/under-provisioned requests, cut node count or cloud cost, size Celery or background workers, tune KEDA listLength or HPA settings, or answer "how many nodes do we need". Works for any namespace with kube-prometheus and cAdvisor metrics.

Deep OWASP reference for security reviews and secure implementation. Covers OWASP Top 10:2025, ASVS 5.0 levels, secure code patterns (injection, auth, error handling, fail-closed), language-specific security quirks for 20 languages, OWASP LLM Top 10 (2025), and Agentic AI security (2026). Use when reviewing code for security vulnerabilities, implementing authentication/authorization, handling user input, hardening an LLM or AI-agent application, or discussing web application security in depth.

Apply a baseline of secure-coding standards to code being written or reviewed. Covers auth, secrets management, input validation, injection prevention, password hashing, PII handling, dependency hygiene, rate limiting, CORS, security headers, HTTPS, error/logging discipline, and infra hardening. Use proactively whenever the user writes or edits authentication/authorization logic, login/session/token code, API endpoints, database queries, file uploads or external input handling, credential sto...

Apply first-principles thinking and design-thinking to product, architecture, and communication decisions. Use this skill whenever the user is framing a problem, choosing between alternatives, debating a tradeoff, questioning a design, adding scope, or saying "should we…" / "that's how it's always done." Also trigger when about to add complexity, recommend a feature, or pick a default — those are where principle drift happens. Explicit triggers include "think first principles", "step back", "...

Give THIS project its own readable brand - after this, every chat card and /report in the project renders in the project's palette, logo, and font. Detects the identity from the repo (design tokens, tailwind config, DESIGN.md, logos) or interviews the user when none exists, then writes a committable .readable/ layer. Use when the user asks to "brand readable", "brand the cards/reports", "make reports match our brand", "رنگ و لوگوی خودمون روی کارت‌ها/گزارش‌ها", or invokes /readable:brand.

Report or change where the readable card server is registered across this machine's Claude desktop profiles. Use FIRST when the user says cards are not rendering, show raw HTML, or stopped working, because status answers that in one command. Also use when they ask to "connect readable", "turn cards on", "disconnect readable", "stop readable writing my config", "کارت‌ها رندر نمی‌شن", "ریدبل رو وصل کن", "ریدبل رو قطع کن", or invoke /readable:connect.

Export a reply or chat card as a standalone styled HTML report file, in the exact readable card template (Persian RTL with Vazirmatn, or English LTR with Inter). Use ONLY when the user explicitly asks to save, export, file, or extend THE CURRENT widget/card ("همین کارت رو ذخیره کن", "کامل‌ترش رو به‌صورت فایل بساز", "save this card as a report") or invokes /readable:report directly. NEVER auto-trigger on generic report-writing requests ("گزارش بساز", "write a report") that do not mention savin...

Turn an answer into a light visual (flowchart, timeline, comparison, architecture sketch), in any language. Trigger ONLY when the user explicitly asks to see something visual, e.g. "show me", "visualize this", "show it as a diagram", "draw it", "با شکل نشون بده", "شماتیکش رو بکش", "تصویریش کن". Never auto-trigger just because content is flow-shaped. Uses the widget tool with minimal, theme-aware SVG.

Deep-clean and simplify a whole repo or codebase, then leave guardrails so it stays clean. Delete dead surfaces, gitignore build artifacts, relocate dev docs, notes, clones and dumps to the project's workspace repo (the assistant-dev home base), split oversized files and functions, unify duplicated logic, draw package boundaries, run a clean-code pass (naming, function size, DRY, KISS, magic numbers, linter-warning burn-down), and encode the rules as architecture tests. Use when the user asks...

Give a repo its own rules at both layers, and put each rule at the cheapest layer that holds it. The told layer is what a project's workspace tells a model before it writes, so it establishes where a session starts, which repo each kind of file belongs in, how the gate is run, the facts nobody can infer, what this repo does not have, the enforced rules as an index rather than a copy, and the handful no script can count. The caught layer is what turns the build red afterwards, as a linter sett...