Jensen-Yao
User个人 Agent 技能库（SKILL.md）：140+ 技能，供 DeepSeek Harness / Claude Code 等 agent 使用
Categories
Indexed Skills (50)
autoresearch
Orchestrates end-to-end autonomous AI research projects using a two-loop architecture. The inner loop runs rapid experiment iterations with clear optimization targets. The outer loop synthesizes results, identifies patterns, and steers research direction. Routes to domain-specific skills for execution, supports continuous agent operation via Claude Code /loop and OpenClaw heartbeat, and produces research presentations and papers. Use when starting a research project, running autonomous experiments, or managing a multi-hypothesis research effort.
evolving-ai-agents
Provides guidance for automatically evolving and optimizing AI agents across any domain using LLM-driven evolution algorithms. Use when building self-improving agents, optimizing agent prompts and skills against benchmarks, or implementing automated agent evaluation loops.
academic-plotting
Generates publication-quality figures for ML papers from research context. Given a paper section or description, extracts system components and relationships to generate architecture diagrams via Gemini. Given experiment results or data, auto-selects chart type and generates data-driven figures via matplotlib/seaborn. Use when creating any figure for a conference paper.
huggingface-accelerate
Simplest distributed training API. 4 lines to add distributed support to any PyTorch script. Unified API for DeepSpeed/FSDP/Megatron/DDP. Automatic device placement, mixed precision (FP16/BF16/FP8). Interactive config, single launch command. HuggingFace ecosystem standard.
audiocraft-audio-generation
PyTorch library for audio generation including text-to-music (MusicGen) and text-to-sound (AudioGen). Use when you need to generate music from text descriptions, create sound effects, or perform melody-conditioned music generation.
autogpt-agents
Autonomous AI agent platform for building and deploying continuous agents. Use when creating visual workflow agents, deploying persistent autonomous agents, or building complex multi-step AI automation systems.
awq-quantization
Activation-aware weight quantization for 4-bit LLM compression with 3x speedup and minimal accuracy loss. Use when deploying large models (7B-70B) on limited GPU memory, when you need faster inference than GPTQ with better accuracy preservation, or for instruction-tuned and multimodal models. MLSys 2024 Best Paper Award winner.
axolotl
Expert guidance for fine-tuning LLMs with Axolotl - YAML configs, 100+ models, LoRA/QLoRA, DPO/KTO/ORPO/GRPO, multimodal support
evaluating-code-models
Evaluates code generation models across HumanEval, MBPP, MultiPL-E, and 15+ benchmarks with pass@k metrics. Use when benchmarking code models, comparing coding abilities, testing multi-language support, or measuring code generation quality. Industry standard from BigCode Project used by HuggingFace leaderboards.
quantizing-models-bitsandbytes
Quantizes LLMs to 8-bit or 4-bit for 50-75% memory reduction with minimal accuracy loss. Use when GPU memory is limited, need to fit larger models, or want faster inference. Supports INT8, NF4, FP4 formats, QLoRA training, and 8-bit optimizers. Works with HuggingFace Transformers.
blip-2-vision-language
Vision-language pre-training framework bridging frozen image encoders and LLMs. Use when you need image captioning, visual question answering, image-text retrieval, or multimodal chat with state-of-the-art zero-shot performance.
brainstorming-research-ideas
Guides researchers through structured ideation frameworks to discover high-impact research directions. Use when exploring new problem spaces, pivoting between projects, or seeking novel angles on existing work.
chroma
Open-source embedding database for AI applications. Store embeddings and metadata, perform vector and full-text search, filter by metadata. Simple 4-function API. Scales from notebooks to production clusters. Use for semantic search, RAG applications, or document retrieval. Best for local development and open-source projects.
clip
OpenAI's model connecting vision and language. Enables zero-shot image classification, image-text matching, and cross-modal retrieval. Trained on 400M image-text pairs. Use for image search, content moderation, or vision-language tasks without fine-tuning. Best for general-purpose image understanding.
constitutional-ai
Anthropic's method for training harmless AI through self-improvement. Two-phase approach - supervised learning with self-critique/revision, then RLAIF (RL from AI Feedback). Use for safety alignment, reducing harmful outputs without human labels. Powers Claude's safety system.
evaluating-cosmos-policy
Evaluates NVIDIA Cosmos Policy on LIBERO and RoboCasa simulation environments. Use when setting up cosmos-policy for robot manipulation evaluation, running headless GPU evaluations with EGL rendering, or profiling inference latency on cluster or local GPU machines.
creative-thinking-for-research
Applies cognitive science frameworks for creative thinking to CS and AI research ideation. Use when seeking genuinely novel research directions by leveraging combinatorial creativity, analogical reasoning, constraint manipulation, and other empirically grounded creative strategies.
crewai-multi-agent
Multi-agent orchestration framework for autonomous AI collaboration. Use when building teams of specialized agents working together on complex tasks, when you need role-based agent collaboration with memory, or for production workflows requiring sequential/hierarchical execution. Built without LangChain dependencies for lean, fast execution.
deepspeed
Expert guidance for distributed training with DeepSpeed - ZeRO optimization stages, pipeline parallelism, FP16/BF16/FP8, 1-bit Adam, sparse attention
drawio-skill
Use when the user requests diagrams, flowcharts, architecture diagrams, ER diagrams, UML / sequence / class diagrams, network topology, ML/DL model figures (Transformer/CNN/LSTM), mind maps, or any visualization. Also use proactively when explaining systems with 3+ components, complex data flows, or relationships that benefit from visual representation. Best suited when the diagram needs custom styling, rich shape vocabulary, swimlanes, or exportable images (PNG/SVG/PDF/JPG). Generates .drawio XML and exports locally via the native draw.io desktop CLI.
dspy
Build complex AI systems with declarative programming, optimize prompts automatically, create modular RAG systems and agents with DSPy - Stanford NLP's framework for systematic LM programming
faiss
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.
optimizing-attention-flash
Optimizes transformer attention with Flash Attention for 2-4x speedup and 10-20x memory reduction. Use when training/running transformers with long sequences (>512 tokens), encountering GPU memory issues with attention, or need faster inference. Supports PyTorch native SDPA, flash-attn library, H100 FP8, and sliding window attention.
gguf-quantization
GGUF format and llama.cpp quantization for efficient CPU/GPU inference. Use when deploying models on consumer hardware, Apple Silicon, or when needing flexible quantization from 2-8 bit without GPU requirements.
gptq
Post-training 4-bit quantization for LLMs with minimal accuracy loss. Use for deploying large models (70B, 405B) on consumer GPUs, when you need 4× memory reduction with <2% perplexity degradation, or for faster inference (3-4× speedup) vs FP16. Integrates with transformers and PEFT for QLoRA fine-tuning.
grpo-rl-training
Expert guidance for GRPO/RL fine-tuning with TRL for reasoning and task-specific model training
guidance
Control LLM output with regex and grammars, guarantee valid JSON/XML/code generation, enforce structured formats, and build multi-step workflows with Guidance - Microsoft Research's constrained generation framework
alibabacloud-find-skills
Use this skill when users want to search, discover, browse, or find Alibaba Cloud (阿里云) agent skills. Triggers include: "find a skill for X", "search alicloud skills", "阿里云有什么 skill","阿里云", "搜索阿里云技能", "有没有管理 ECS/RDS/OSS 的 skill", "阿里云 skills 有哪些类目", "帮我找一个 skill", "browse alicloud skills", "list alicloud skill categories", "is there an alicloud skill that can...", "what alicloud skills are available", "XX Skill 的内容是什么", "我想了解阿里云 XX Skill 具体做什么","帮我安装阿里云 Skill","使用阿里云相关的skill", "阿里云 agent skill 市场", "搜一下阿里云的 skill", "建一个数据分析项目有没有相关 skill".
animation-vocabulary
Reverse-lookup glossary that turns a vague description of a web animation or motion effect into its exact term ("the bouncy thing when a popover opens" → Pop in; "the iOS rubber-band scroll" → Rubber-banding). Use when the user asks "what's it called when…", or describes a motion effect without knowing its name and wants the right word to prompt an AI or designer with. For naming an effect, not designing or building one.
app-shell-ui
Build or restyle frontends using App Shell UI — a clean, restrained macOS-style desktop-tool shell (left nav + content, soft surfaces, raised cards, single brand accent, settings-list language) with first-class light and dark themes. Use when the user asks for app-shell-ui, /app-shell-ui, 桌面工具风 UI, 系统设置感界面, 应用外壳, sidebar shell, Clash/设置风格前端, light/dark desktop client UI, 深色主题, 浅色主题, theme toggle, or wants pages that share one reusable app chrome across chat / settings / console / mail-like layouts. Also use for high-fidelity HTML/React demos from tokens and component recipes, or when the user asks to reuse the bundled Top 50 React UI component library / gallery.
apple-design
Apple's approach to interface design and fluid, physical motion, translated for the web. Use when building or reviewing gesture-driven UI, spring animations, drag/swipe/sheet interactions, momentum and interruptible transitions, translucent materials and depth, typography (optical sizing, tracking, leading), reduced-motion, or the design foundations (feedback, spatial consistency, restraint) behind Apple-style interfaces.
bugbounty-workflow
授权范围内(补天/SRC 等赏金平台)漏洞挖掘全流程工作流:侦察→挖掘→验证→报告。当用户要��"挖洞""挖漏洞""测目标""渗透测试""打点""提交SRC/补天"时使用。
cli-anything
Use when the user wants Codex to build, refine, test, validate, or list CLI-Anything harnesses for GUI applications or source repositories. Adapts the full CLI-Anything methodology to Codex without changing the generated Python harness format.
cli-creator
Build a composable CLI for Codex from API docs, an OpenAPI spec, existing curl examples, an SDK, a web app, an admin tool, or a local script. Use when the user wants Codex to create a command-line tool that can run from any repo, expose composable read/write commands, return stable JSON, manage auth, and pair with a companion skill.
cnki-advanced-search
Perform advanced search on CNKI with field filters like author, title, journal, date range, source category (SCI/EI/CSSCI/北大核心). Use when user needs precise filtered search beyond simple keywords.
cnki-download
Download a paper PDF/CAJ from CNKI. Requires user to be logged in. Use when user wants to download a specific paper.
cnki-export
Export paper from CNKI and push to Zotero, or save as RIS file. Use when user wants to save a paper to Zotero or export citation data.
cnki-journal-index
Query journal indexing/inclusion status on CNKI - check which databases include a journal (北大核心, CSSCI, CSCD, SCI, EI, etc.), get impact factors and evaluation data. Use when user asks about a journal's level, indexing, or ranking.
cnki-journal-search
Search for journals/publications on CNKI by name, ISSN, CN, or sponsor. Use when the user wants to find a specific journal or browse publications.
cnki-journal-toc
Browse journal issues, view table of contents, and download original TOC PDF from CNKI. Use when user wants to see papers in a specific journal issue or download the original cover/TOC pages.
cnki-navigate-pages
Navigate CNKI search result pages (next/previous/specific page) or change sort order. Use when user wants to see more results or change sorting.
cnki-paper-detail
Extract full paper details from a CNKI paper page including title, authors, affiliations, abstract, keywords, fund, classification. Use when the user needs detailed information about a specific paper.
cnki-parse-results
Parse current CNKI search results page into structured paper data (title, authors, journal, date, citations). Use after a search has been performed and you need to extract the results.
cnki-researcher
Coordinate CNKI research workflows in Codex using Chrome DevTools MCP. Use when the user asks to search CNKI papers, inspect paper details, browse journals or issues, check journal indexing, download CNKI files, export citations, handle institution/off-campus CNKI login, or combine these tasks into a literature research workflow.
cnki-search
Search CNKI (中国知网) for papers by keyword. Use when the user wants to find academic papers on a topic.
dramaclaw
Operate, inspect, configure, troubleshoot, and continue DramaClaw projects according to the official workflow. Use whenever the user mentions DramaClaw, NovelVideo, 虾导, 虾料, 虾塘, 虾镜, 虾画, 虾格, 虾条, CLI/API/Web project synchronization, model or storage configuration, characters, identities, portraits, scenes, props, voices, episodes, Beats, scripts, sketches, first frames, audio, video, composition, project status, resume/continue, or asks to generate/fix a DramaClaw production. Enforces asset, image, and review gates before downstream generation.
emil-design-eng
This skill encodes Emil Kowalski's philosophy on UI polish, component design, animation decisions, and the invisible details that make software feel great.
find-animation-opportunities
Search a codebase or UI for places that don't animate but should, and reject everything that shouldn't. Read-only; it proposes motion with exact values, it does not implement it. Use when the user asks "what could be animated here?" or wants to "make this feel more alive". For fixing existing animations, use improve-animations or review-animations instead.
hatch-pet
Create, repair, validate, visually QA, and package Codex-compatible animated pets and pet spritesheets from character art, generated images, company or prospect brand cues, or visual references. Use when a user wants a lightweight-worker Codex pet workflow, a non-pixel custom pet style, a prospect or company mascot pet, or a full 8x9 animated pet atlas with transparent unused cells, QA contact sheets, and pet.json packaging. This skill composes the installed $imagegen system skill for visual generation and uses bundled scripts for deterministic spritesheet assembly.
chinese-plot-labels
Use when Codex writes or modifies Python, Java, or C++ code that generates charts, plots, figures, or saved image outputs. If the user has not explicitly specified another language, require Chinese titles, axis labels, and legends in generated result images.
Bio shown is the top-scored skill's repo description as a fallback — real GitHub bios land in a future update.