Machine learning

用 Playwright 从深知晓、深知晓（深度溯源）、豆包、腾讯元宝、DeepSeek 和通义千问网页端无损采集完整回答、原始引用、引用标记和现场存证。用于 Fact-Check-X 的原始答案采集阶段，或任何需要比较多个 AI 网页回答与参考来源的任务。

Karpathy's LLM Wiki: build/query interlinked markdown KB.

First-principles ML: classical methods, optimization, Gaussian processes, deep learning, probabilistic inference, generative models, RL.

Use when the user wants to test, compare, or choose local Ollama models for their machine.

Detect command-and-control (C2) traffic tunneled over DNS from tools like Iodine, dnscat2, dns2tcp, and Cobalt Strike DNS beacon, using Shannon entropy…

Answer Engine Optimization (AEO): optimize content to be cited by LLMs (ChatGPT, Claude, Perplexity, Gemini) in their answers.

Context management engine for AI coding agents.

Data science across machine learning, statistical modeling, and experimentation.

MLOps across model deployment, ML pipelines, monitoring, and feature stores.

Design RAG pipelines: chunking, retrieval evaluation, and architecture.

Computer vision engineering for object detection, segmentation, and visual AI, covering CNN and Vision Transformer architectures and ONNX/TensorRT deployment.

Prompt engineering and LLM evaluation.

中文去 AI 腔一整套（装完即用，零 API Key —— 装了本 skill 的 agent 自己就是那个 LLM）。 六个流程：① 生成即去 AI —— 写中文时按本 skill 的写作指南落笔，初稿就不带 AI 腔； ② 复查改写 —— 产出后对照本 skill 的清单再过一遍，定点拆模板句式； ③…

ML engineering skill for productionizing models, building MLOps pipelines, and integrating LLMs.

This skill should be used when a user reports an issue with OpenHands Enterprise (OHE) on a self-hosted (Replicated VM-based) installation.

Optimize content to rank in AI search engines (AI Overviews, Perplexity, ChatGPT) via generative engine optimization (GEO), citability audits, and schema…

Wire an application to heylook (heylookitsanllm), a local multimodal LLM server on Apple Silicon serving MLX and gguf models over one Anthropic…

Parse, typeset, explain, and evaluate research-paper equations, notation, symbols, derivations, losses, objectives, tensor shapes, assumptions, and…

Threadlight evals leg: offline batch quality, Foundry Continuous Evaluation and champion-challenger gates.

Collect and compare numerical results across hardware and software stacks for inference, SFT (full/LoRA) and RL.

Evaluate a hardware accelerator's integration capability with PyTorch.

Guided workflow for taking any ML codebase — including one with no experiment tracking at all, or one full of MLflow 2-era idioms — to a production-grade…

使用自定义大模型（本地 Ollama 模型或远程 OpenAI 兼容 API，如 DeepSeek/OpenRouter）处理工作内容。当用户要求"用自定义大模型/自定义模型/本地模型/Ollama/DeepSeek/OpenRouter/qwen3…

用于诊断、评审或评分整份简历。当用户要求「诊断简历」「简历有什么问题」「看看简历哪里要改」「简历打分」「评审简历」「resume review」，或上传 PDF/DOCX 并要求分析问题、给出评分或修改建议时触发。输出 3 类诊断（NO_DATA / NO_FILL / DESCRIPTION_SUGGEST）和…