Create academic presentation slide decks and optionally demo videos from research papers. Use when the user asks to "make slides", "create a deck", "make a presentation", "demo video", "paper slides", "conference talk slides", or wants to turn a paper into a visual presentation. Covers slide generation, narration scripts, TTS audio, and video assembly.

Trains large language models (2B-462B parameters) using NVIDIA Megatron-Core with advanced parallelism strategies. Use when training models >1B parameters, need maximum GPU efficiency (47% MFU on H100), or require tensor/pipeline/sequence/context/expert parallelism. Production-ready framework used for Nemotron, LLaMA, DeepSeek.

Search published venue papers (IEEE, ACM, Springer, etc.) via Semantic Scholar API. Complements /aris-arxiv (preprints) with citation counts, venue metadata, and TLDR. Use when user says "search semantic scholar", "find IEEE papers", "find journal papers", "venue papers", "citation search", or wants published literature beyond arXiv preprints.

Draft LaTeX paper section by section from an outline. Use when user says "写论文", "write paper", "draft LaTeX", "开始写", or wants to generate LaTeX content from a paper plan.

Mandatory pre-flight compute resource check before running experiments. Detects whether local/remote GPU or compute resources are actually available. If resources are unavailable, STOPS the experiment pipeline immediately and reports to the user — preventing the model from hallucinating fake experiment results. Use when: about to run experiments, deploy training, or any GPU-intensive task.

Search, download, and summarize academic papers from arXiv. Use when user says "search arxiv", "download paper", "fetch arxiv", "arxiv search", "get paper pdf", or wants to find and save papers from arXiv to the local paper library.

Discover top LinkedIn influencers and voices by topic, industry, follower count, and country. Use when you need to find the top 100 voices in a space, build influencer lists for outreach, or identify thought leaders on LinkedIn.

Scrapes LinkedIn job postings using the JobSpy library (python-jobspy). Use this skill whenever the user wants to find jobs on LinkedIn, search for open roles, pull job listings, build a job pipeline, source job targets for GTM research, or monitor hiring signals. Even if the user just says "find me some jobs" or "what roles is [company] hiring for", use this skill. It runs a local Python script that outputs a CSV of job postings with title, company, location, salary, job type, description, and direct URLs.

Extract competitor and customer intelligence from any company's landing page HTML. Discovers tech stack, analytics tools, ad pixels, customer logos, SEO metadata, CTAs, hidden elements, and more. No API keys required.

Render a punchy ~12s vertical (9:16) music-only direct-response OFFER ad as a 4-beat kinetic-typography film — HEADLINE slam → real PRODUCT drop → CLAIM/proof → CTA pill — from one config of copy slots, a real product photo, a brand palette, fonts, bpm, and beat split. DETERMINISTIC + FREE (a bundled Remotion project; springs + interpolate, no AI-gen for visuals). Backgrounds are engine gradient divs off the palette, props are inline SVG, the ONLY composited bitmap is the REAL product photo (objectFit:contain, never stretched), and ALL headline/claim/CTA/URL/wordmark text is typeset in the engine — never AI-rendered (the format's credibility guard). A driver binds the config to Remotion input props, renders the 9:16 master, and derives a 1:1 center-crop with ffmpeg. Two gating checks run before render (claim verbs must match the product's physical format; the claim beat needs an edge-entry mechanism prop). Use for the motion-graphics-offer-ad format.

Use this skill when you develop, design, build, deploy, debug, migrate, or troubleshoot CloudBase (腾讯云开发, 云开发, TCB, 微信云开发) projects. Covers Web, 微信小程序, 小程序, uni-app, mobile (iOS, Android, Flutter, React Native). UI (页面, 界面, 表单, form, dashboard, prototype, 原型); auth (登录, 注册, OAuth, 微信登录, publishable key); databases (NoSQL 文档数据库, MySQL 关系型数据库, PostgreSQL/CloudBase PG, app.rdb(), queryPgDatabase/managePgDatabase, CRUD, 查询, security rules); 云函数/cloud functions (serverless, scf_bootstrap); CloudRun (云托管, Dockerfile); 云存储. Built-in AI (内置大模型, AI 对话, streaming, 流式输出, 图片生成, generateText, streamText, createModel, generateImage, TokenHub, Hunyuan, hunyuan-exp, DeepSeek, deepseek, GLM, Kimi, Token Credits 资源包, 小程序成长计划). 第三方大模型, 大模型接入, 大模型调用, LLM API, AI agent, 智能体, AI Agent, AG-UI, LangGraph. Ops (巡检, 诊断, health check, 日志, troubleshooting). Spec (需求文档, 技术方案, requirements, tasks.md). Do NOT use for non-CloudBase projects, pure frontend without CloudBase, or self-hosted backends without CloudBase.

Use when users need to implement, integrate, debug, build, deploy, or validate a Web frontend after the product direction is already clear, especially for React, Vue, Vite, browser flows, or CloudBase Web integration.

[Deprecated] Use when building frontend Web apps that talk to CloudBase Relational Database via @cloudbase/js-sdk – provides the canonical init pattern so you can then use Supabase-style queries from the browser. New environments should use PostgreSQL with app.rdb() — see postgresql-development skill instead.

Use when the user wants to develop, run, preview, save, deploy, or roll back a CloudBase Web app in this conversation as a Lovable/Codex-Sites-like vibe- coding session — i.e. any request that maps to "spin up a React+Vite project, see it live in the browser, iterate, save versions, deploy". Activate ONLY for browser-rendered Web projects based on the CloudBase official React (or Vue) + Vite template. Do NOT activate for: WeChat/Alipay mini-program, React Native / Flutter / Electron, pure cloud functions / CloudRun backends, or Next.js / Nuxt / Astro / Remix / SvelteKit (those frameworks are explicitly out of scope).

CloudBase Run backend development rules (Function mode/Container mode). Use this skill when deploying backend services that require long connections, multi-language support, custom environments, AI agent development, or migrating existing/GitHub apps that need VPC access to MySQL/PostgreSQL/Redis. Also use when diagnosing CloudRun container deploy failures (deploy_failed, readiness/probe failed, image won't start, docker.io pull loops). For stateless HTTP services, prefer HTTP cloud functions.

Validates equivalence between LiteRT models (litert_lm) and PyTorch models (transformers). Use when you need to verify that an exported LiteRT model produces the same outputs as the original Hugging Face model. Supports multi-turn conversations and custom prompts.

Use when the user wants to run XCUITests, parse test results, view test failures, or export test attachments.

Use this agent to scan Swift code for camera, video, and audio capture issues including deprecated APIs, missing interruption handlers, threading violations, and permission anti-patterns.

Use when building a managed team skills library for a real stack. Map work to shelves, browse before curating, write meaningful `whyHere` notes, and create a starter pack once the first pass is solid.

Examines a specific product's detailed page to verify it matches the user's requirements, checking price, description, features, and reviews. Use when you have navigated to a product detail page from search results and need to confirm the product meets all user-specified constraints before purchasing. It provides a final suitability assessment with a clear proceed-or-reject recommendation.