Loading...
Loading...
Found 87 Skills
Search and recommend Claude Code skills from trusted marketplaces
Explore and analyze GitHub repositories related to a research topic. Reads deep-research output, discovers repos from multiple sources, deeply analyzes code, and produces integration blueprints.
Skill for researching best practices. Triggered when you need to understand methodologies, tools, and best practices in a specific field. Trigger words: research, learn about, what methods are there, best practices.
Write publication-ready ML/AI papers for NeurIPS, ICML, ICLR, ACL, AAAI, COLM. Use when drafting papers from research repos, structuring arguments, verifying citations, or preparing camera-ready submissions. Includes LaTeX templates, reviewer guidelines, and citation verification workflows.
Adds PyTorch FSDP2 (fully_shard) to training scripts with correct init, sharding, mixed precision/offload config, and distributed checkpointing. Use when models exceed single-GPU memory or when you need DTensor-based sharding with DeviceMesh.
Provides guidance for PyTorch-native agentic RL using torchforge, Meta's library separating infra from algorithms. Use when you want clean RL abstractions, easy algorithm experimentation, or scalable training with Monarch and TorchTitan.
Meta's 86M prompt injection and jailbreak detector. Filters malicious prompts and third-party data for LLM apps. 99%+ TPR, <1% FPR. Fast (<2ms GPU). Multilingual (8 languages). Deploy with HuggingFace or batch processing for RAG security.
Facebook's library for efficient similarity search and clustering of dense vectors. Supports billions of vectors, GPU acceleration, and various index types (Flat, IVF, HNSW). Use for fast k-NN search, large-scale vector retrieval, or when you need pure similarity search without metadata. Best for high-performance applications.
USE FOR AI-grounded answers via OpenAI-compatible /chat/completions. Two modes: single-search (fast) or deep research (enable_research=true, thorough multi-search). Streaming/blocking. Citations.
USE FOR query autocomplete/suggestions. Fast (<100ms). Returns suggested queries as user types. Supports rich suggestions with entity info. Typo-resilient.
Robust URL-to-Markdown extraction for OpenClaw workflows. Use this when the user needs to "extract/summarize/convert a webpage to Markdown" (especially for WeChat official accounts at mp.weixin.qq.com) and web_fetch or browser access is blocked or returns messy content. It first uses a low-cost probe via web_fetch, then falls back to the official MinerU API (through the local mineru-extract skill), and returns a traceable result contract with source links.
Scalable data processing for ML workloads. Streaming execution across CPU/GPU, supports Parquet/CSV/JSON/images. Integrates with Ray Train, PyTorch, TensorFlow. Scales from single machine to 100s of nodes. Use for batch inference, data preprocessing, multi-modal data loading, or distributed ETL pipelines.